Debt Challenges Ahead

Carmen M. Reinhart

Harvard University

The World Bank, June 1, 2020

1

Roadmap

• Pre-pandemic vulnerabilities and recent developments

• Hidden debts: What to expect in a crisis

• Debt build-ups and debt management during the crisis

• Temporary debt payment suspension and other options

• Dealing with debt overhangs and debt restructuring

• Final thoughts on the coming banking crises

Reinhart 2

Pre-pandemic vulnerabilities in historical context and recent developmentspublic debtexternal debt at the time of defaultthe VIX (risk aversion) capital flows to EMscredit rating downgradescost of financing, yield ratios

Reinhart 3

Public debt and major crises: 45 emerging and developing countries, 1900-2020

Reinhart 4

10

20

30

40

50

60

70

80

90

100

1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020

% o

f GDP

SPre GFC, 2006-2008Deleveraging2008 avg. 37.6%

Pre-depression default wave,1929-1931Debt build-up1931 avg. 47.2%

Pre COVID-19.2017-2019Debt build-up2019 avg. 55.9%

Pre 1980s debt crisis,1980-1982Debt build-up1982 avg. 45.8%

Note: WEO 2020 estimates at this stage are likely to underestimate the debt spike this year. Sources: Reinhart and Rogoff (2009) and sources cited therein, World Economic Outlook (2020).

5

Defaults and restructurings at low levels of external debt by advanced economy standards are commonplace in EMsFrequency distribution of external debt ratios in middle-income countries at the time of default 1970-2001 (Reinhart, Rogoff, and Savastano, 2003)

This fact holds when updated to include episodes with recent cresit events.

Reinhart

Weekly Net Non-Resident Purchases of EM Stocks and Bonds ($ billion) (Excludes Turkey and Mexico)

Reinhart 6

-100

-50

0

50

100

150

-30

-25

-20

-15

-10

-5

0

5

10

15

Weekly, blue(left scale)

52-week sum,Red (right scale)

It took 53 weeks during the Global Financial Crisis to get to the cumulative level of outflows recorded in 6 weeks

Source: IIF (2020)

Reinhart 7

Historically capital flows to developing counties have been negatively correlated with the global share (incidence) of countries in default (Reinhart, Reinhart and Trebesch, 2020). The surging odds of defaults evident in credit rating changes certainly map onto the sudden stop in capital flows to EMs

7

Share of downward revised sovereign outlooks by major rating agencies, 3-month sums1990:1-2020:5

Number of downgraded sovereigns,3-month sums1980:1-2020:5

0.00

5.00

10.00

15.00

20.00

25.00

30.00

Jan

19

80

May

19

81

Sep

19

82

Jan

19

84

May

19

85

Sep

19

86

Jan 1

988

May

19

89

Sep

19

90

Jan 1

992

May

19

93

Sep

19

94

Jan

19

96

May

199

7

Sep

19

98

Jan

20

00

May

200

1

Sep

20

02

Jan

20

04

May

20

05

Sep

20

06

Jan

20

08

May

20

09

Sep

20

10

Jan 2

012

May

20

13

Sep

20

14

Jan 2

016

May

20

17

Sep

20

18

Jan

20

20

3 month sums

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

Jan

19

90

Jan

19

91

Jan

19

92

Jan

19

93

Jan

19

94

Jan

19

95

Jan

19

96

Jan

19

97

Jan

19

98

Jan

19

99

Jan

20

00

Jan

20

01

Jan

20

02

Jan

20

03

Jan

20

04

Jan

20

05

Jan

20

06

Jan

20

07

Jan

20

08

Jan

20

09

Jan

20

10

Jan

20

11

Jan

20

12

Jan

20

13

Jan

20

14

Jan

20

15

Jan

20

16

Jan

20

17

Jan

20

18

Jan 2

01

9

Jan 2

02

03 month sums

Sources: Fitch, Moody’s, Standard & Poors, and Trading Economics.

Reinhart 8

Risk aversion is also a key factor in the sudden stop in capital flowsVolatility and Risk Aversion CBOE VIX and VIX proxy:1885:2-2020:5

8

0.00

0.20

0.40

0.60

0.80

1.00

1.20

02

-18

85

01

-18

88

12

-18

90

11

-18

93

10

-18

96

09

-18

99

08

-19

02

07

-19

05

06

-19

08

05

-19

11

04

-19

14

03

-19

17

02

-19

20

01

-19

23

12

-19

25

11

-19

28

10

-19

31

09

-19

34

08

-19

37

07

-19

40

06

-19

43

05

-19

46

04

-19

49

03

-19

52

02

-19

55

01

-19

58

12

-19

60

11

-19

63

10

-19

66

09

-19

69

08

-19

72

07

-19

75

06

-19

78

05

-19

81

04

-19

84

03

-19

87

02

-19

90

01

-19

93

12

-19

95

11

-19

98

10

-20

01

09

-20

04

08

-20

07

07

-20

10

06

-20

13

05

-20

16

04

-20

19

RRT VIX proxy CBOE VIX Top 1% readings

Top 1% Monthly Readings

for VIX proxy, 1885:2 - 2020:5

1929: Oct - Nov

1931: Oct

1932: Feb, Jun, Aug - Oct

1933: Mar, Jul, Oct

1937: Oct

1987: Oct

2008: Oct - Nov

2020: Mar

Sources: Reinhart, Reinhart, and Trebesch (2020) based on Schwert (1990), Thomson Reuters EIKON (2019), FRED.Note: Correlation of CBOE VIX and RRT VIX proxy, 1990-2020 is 0.89. VIX proxies for UK and US were also constructed at an annual frequency for the full 1815-2018 sample.

Top 1 percentile

The March COVID-19VIX proxy spike is thelargest

on record

For comparisons over time (1995-2020) of yields of developing country debt to the US risk-free rate when there is a trend in interest rates, spreads are deceptive…This highlights how comparatively costly it is for developing countries to raise new funds or roll over existing debts.

Reinhart 9

0

2

4

6

8

10

12

14

16

18

20

10/28/1995 4/19/2001 10/10/2006 4/1/2012 9/22/2017 3/15/2023

JPM EMBI+ yield relative to the five-year Treasury yieldpercentage points and ratio

Spread Ratio

Yield ratiosprovide a betterindication of relative risk

Source: Bloomberg, May 31, 2020.

Hidden debts: What to expect

definition of the official sectormismeasurementarrears contingent liabilities central bank reserves: gross versus net

Reinhart 10

11

Central bank debt: To this day, even when the numbers are published, these are not included as part of general government debt (ie, Argentina’s short-term Lebacs). In the event of Euro-area exits, Target2 balances (currently running at around 20-40% of GDP for Greece, Italy, Portugal and Spain) are external central bank debt. Unseen: In June 1997, the Thai finance minister ‘discovered’ that the Bank of Thailand had already spent US$ 28 billion out of US$ 30 billion of its international reserves in the course of forward market interventions to defend the baht.

Nonsecuritized, floating debt (arrears): Unpaid bills to suppliers and, in more desperate cases (Russia 1998) unpaid pensions and wages to public sector employees. PEMEX is a more recent example of unpaid wage bills.

Misreporting and other off-balance sheet: Greece-Goldman Sachs debt swaps: Greek dollar and yen-denominated debt was swapped at historical euro exchange rates to cosmetically reduce the overall level of debt.

External debts to China are, at best, partially incorporated in the government accounts of many low and middle-low income developing countries (Horn, Reinhart, and Trebesch, 2019).

Offshore derivative operations of banks: These can leverage banks’ holdings of government debt (a significant hidden debt problem during the Mexican banking/peso crisis of 1994-1995). With Mexican bonds (Tesobonos) as collateral, Mexican banks took on short-term dollar debt, that was for the most part unhedged. As the value of the collateral sank, margin calls increased along with rollover risk.

Implicit guarantees and moral hazard:

Puerto Rico (PR) began issuing “Appropriation bonds” in 2000 indirectly through government-owned entities and made repayment contingent upon the Legislature’s appropriating funds for this purpose. These bonds were not counted as debt under the debt limit. Yet PR appropriation debt was, for practical purposes, guaranteed by the government and charged to its taxpayers.

Private sector debt Especially external debt of banks (Diaz Alejandro, 1985) can overwhelm an otherwise healthy fiscal situation (Chile 1981, Iceland, Ireland, Spain, 2007-2008); corporate debt (Korea and Indonesia, 1997).

The extent of the contingent liability problem for corporate debt may itself be underestimated. Coppola, Maggiore, Neiman, and Schreger (2020) highlight the role of corporate debt issued in tax havens not reported in country’s official debt statistics.

Hidden debts are unpleasant surprises that when revealed have often undermined the credibility of existing safety nets and set in motion runs

Reinhart

The (very) old problem of contingent liabilities; public debt was low, but the private sector was highly indebted (domestic and external)Source: Reinhart (2015)

The debt surges reflect the effects of deep recessions (revenue collapse and higher spending) and the bail-out of the financial sector

Percent

General Government debt/

GDP in 2007

Iceland 27.1

Ireland 24.0

0

20

40

60

80

100

120

140

1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Iceland

Ireland

Reinhart 12

External debt and estimates of “hidden debts” to China

13

13

Note the

scales

Reinhart

Horn, Reinhart, and Trebesch (2019, and forthcoming)

Debt buildups and debt management during crises

Build-up in domestic and external debt as the crisis nearsDebt management issues:

the bias to short maturities and roll-over riskco-ordination with the central bank (big buyer of debt)currency of denominationthe private sector composition

Reinhart 14

The run-up in domestic and external debt on the eve of external default: Eighty-nine episodes, 1827-2003 Source: Reinhart and Rogoff (2009)

Reinhart 15

90.00

100.00

110.00

120.00

130.00

140.00

150.00

t-4 t-3 t-2 t-1 T

External Domestic

t-4=100

Default year

This is along the lines of what we are seeing in 2020.More of the action is in domestic debt,which is often harder to track

Is there a problem with this strategy? The need for central bank and treasury co-ordination on debt management

Central Bank Chief Roberto Campos Neto has said that his preferred version of QE would flatten the yield curve without expanding the monetary base or the bank’s balance sheet, by buying long-dated bonds while selling short-dated debt - akin to the Fed’s “Operation Twist” in 2011.https://www.reuters.com/article/us-brazil-economy-qe-analysis/brazil-set-to-start-qe-cautiously-but-may-need-to-bring-out-bazooka-idUSKBN2221G1

Source: Reuters, April 20, 2020

Reinhart 16

Reinhart 17

Financial fragility and short-term debt surges on the eve of crisesAmong the manifestations of fragility is the change in the maturity composition of the debt. Short-term debts escalate on the eve of banking crises; the ratio of short-term to total debt about doubles from 12 to 24 percent. A similar pattern emerges in the run-up to sovereign defaults. `

Whether the rise in short-term debt reflects growing perceived risk and reluctance by lenders to extend longer term debt, (Broner, Lorenzoni, and Schmukler, 2013 present evidence of the sharp steepening of the yield curve prior to EM crises), or the hope of borrowers that better times and borrowing terms are around the corner (thus underestimating roll-over risks), a higher short-term debt ratio exposes a country to greater risk of a self-fulfilling panics.

.

Source: Reinhart and Rogoff (2011).

Share of Short-term Gross External Debt(Public plus Private): Emerging Markets, 1970–2009 (in percent)

Banking crises (black bars)

Temporary debt payment suspension and other optionsThe debtors

The creditors: Where we are

Relative merits of standstills

Reinhart 18

EMs and frontier economies are in the high-yield universe. Significant segments of that market have official support since the onset of the crisis

Reinhart 19

Federal ReserveUS high-grade

corporates, fallen angels and other lower-rated

corporates

Federal ReserveMunis and sub-

sovereigns

G-20 bilateral creditors

IDA countries

Non-IDA eligible EMs and the lowest-rated US corporates are out herewithout official support.

6-month debtpayment moratorium

Some of the borrowing fromFed’s credit facilities

offer grace periodof 6 months

Federal ReserveMain StreetFacility with

banks

The debtors:• Current G-20 initiative is limited to IDA countries and participation is

on a voluntary basis.

• Take-up by IDA countries has been surprisingly low. Fear of credit rating downgrades has been cited as a reason.

• Fear of greater scrutiny of public accounts that the G-20 initiative includes may be another unspoken reason (remember our hidden debt discussion)

• List of developing nations affected by COVID-19 is much, much longer than the 76 IDA countries

• Within that longer list, several small, tourism-based economies with no capital market access do not qualify for IDA

• Increased risk aversion on the part of investors and credit rating downgrades (as discussed earlier) make borrowing much costlier for emerging markets at a time when financing needs have increased markedly

• Distance to default is shrinking rapidly for many economies.

Reinhart 20

The creditors: where we are

• Among official bilateral creditors, there is the uncertainty of whether and how China (very important, see chart), Saudi Arabia, and other creditors outside the Paris Club will implement the G-20 standstill. Recently, there are some encouraging signs from the Chinese government on that score.

• Among some of the IFIs (WB included), there is the concern that participation on a standstill would lead to a weakening of the Preferred Creditor Treatment (PCT) possibly jeopardizing the triple A rating of these institutions, which would raise the cost of funding and therefore lending.

• Backpaddling (see Bolton, et. al, 2020): As to the private sector, historically voluntary participation is complicated and often drawn out (more on this later) and more likely to be done on a country-by-country basis.

• A more general approach (see the Bolton Committee) invokes a doctrine of customary international law known as the "state of necessity." The doctrine itself is straightforward -- if a country must breach its international commitments in order to deal with an existential threat (war is the usual culprit), its obligations will be deemed suspended. Commentators argue about whether necessity applies to a state's contractual commitments to private parties or only to its treaty obligations.

Reinhart 21

China is the most important official creditor of developing and emerging countries: It’s participation in any debt relief initiative is key.

Reinhart, 22

0

50

100

150

200

250

300

350

400

2000 2003 2006 2009 2012 2015

Agg

rega

te p

ub

lic d

ebt

in b

illio

ns

of

USD

Debt to China Debt to all Paris Club governments

Debt to IMF Debt to World Bank

Source: Horn, Reinhart, and Trebesch (2020).

On debt service standstills• Standstills are NOT a substitute for debt restructuring if debt sustainability

is in question. More of this to follow.

• No historical antecedent to compare to. The 1931 Hoover Moratorium, which lasted 1 year applied only to official WWI debts. The amounts owed were substantial (see Reinhart and Trebesch. 2016) but only applied to about 15 countries.

• Debt servicing burdens are high in many developing and emerging markets (for IDA about 12% of revenues in 2019 and rapidly climbing).

• It limits the immediate debt build-up associated with coping with the pandemic. Many countries were already facing debt problems.

• Can free up resources quickly in an emergency.

• Can provide time to develop a more sustainable debt strategy—if debt restructuring is necessary

• Raises concerns about stigma and credit ratings (even for countries that do not choose not to participate)

• At the moment, it does not help countries which relied on multilaterals and private creditors (even among the IDA countries)

• As noted, it does not help countries that do not meet IDA low-income criteria and this is a GLOBAL crisis.

Reinhart 23

Other options to provide debt relief (support) beyond new rapid and highly concessional lending by IFIs (these are not mutually exclusive)

• Should multilaterals participate in the standstill? As noted, there is the concern that participation on a standstill would lead to a weakening of the PCT possibly jeopardizing the triple A rating of IFIs, which would raise the cost of funding and therefore lending. But failure to do so raises credibility issues with private creditors and some governments the “Do as I say but not as I do” problem.

• An IFI facility to buy sovereign external debt (modeled after what major central banks have done). Reduces some the recent distortions in the high-yield market, would reduce odds of fire sales, and provide a more stable creditor base (can reduce rollover risks). Problem is scale, as IFIs are capital-constrained.

• Should IFIs lobby to have the major central banks buy their debts (they are buying far riskier assets already)? This would provide support for fund-raising at a critical time.

• Credit rating agencies are notoriously procyclical; IFIs were designed to be countercyclical (relative to private lending). Should credit rating agencies exert such influence over these institutions?

Reinhart 24

Undoing debt overhangs with a focus on debt restructuringMenu of optionsWhat to expect from debt restructuring

Magnitude of haircutsLength of default spells and number of negotiationsMagnitude of debt relief (and type)Re-access to external capital and aftermath

Reinhart 25

Undoing debt overhangs: Menu of options

• Debt/GDP ratios have been reduced by:

• (i) economic growth;

• (ii) fiscal adjustment/austerity;

• (iii) explicit default or restructuring;

• (iv) a sudden surprise burst in inflation; and

• (v) a steady dosage of financial repression that is accompanied by an equally steady dosage of inflation.

• (Options (iv) and (v) are only viable for domestic-currency debts).

26

Renhart

This is what I am going to focus on

Reinhart, Reinhart, Trebesch 27

BRA

ITA

JPN

MEX

PER

POL

ROU

URY

SRB

BRA

BGR

CHL

DOMMEX

NGA

PAK

PAN

POL

POL

RUS

TUR

COL

TUR

UKR

SRB

CHN

POL

CRI

CZK

ECU

GRC

GTM

HUNALB

DZA

DZA

ARG

ARG

ARG

ARG

ARG

ARG

ARG

AUT

AUT

BLZ BLZ

BOL

BOL

BOL

BOL

BIH

BRABRA

BRA

BRA

BRABRA

BRA

BGR

BGR

CMR

CHL

CHL

CHL

CHL

CHL

CHL

CHL

CHN

COL

COL

COL

COL

COL

COD

COG

COD

CODCOD

COD

COD

COD

COG

CRI

CRI

CRI

CRICRI

CRI

CRI

CRI

CRI

CRI

CIV

CIV

CIV

HRV

CUB

CUBCUB

CUB

CUB

CZK

DOM DOM

DOM

DMA

DOM

DOM

DOM

DOM

ECU

ECU

ECU

ECU

ECU

ECU

ECU

ECU

ECU

ECU

ECU

EGY

EGY

EGY

SLV

SLVSLV

SLV

SLV

SLV

EST

ETH

FIN

GAB

GAB

GMB

DEU

DEU

GRC

GRC

GRC

GRC

GRD

GRD

GTM

GTM

GTM

GTM

GTM

GIN

GINGUY

GUY

HND

HND

HND

HND

HUN

IRQ

JAM

JAM

JAM

JAM

JAM

JAM

JAM

JOR

KEN

LVA

LBR

LBR

LBR

LBR

LBR

LTU

MKD

MDG

MDG

MDG

MDG

MWI

MWI

MRT

MEX

MEX

MEX

MEX

MEX

MEX

MEX

MEX

MEX

MEX

MEX

MEX

MEX

MDA

MDA

MARMAR

MAR

MOZMOZ

NIC

NIC

NIC

NIC

NIC

NIC

NIC

NIC

NIC

NIC

NIC

NGA

NER

NER

NGA

NER

NGA

NGA

NGA

PAN

PAN

PRY

PRY

PRY

PRY

PRY

PER

PER

PER

PER

PER

PER

PER

PER

PHLPHL

PHL

PHL

POL

POL

POL

POL

PRT

PRT

PRT

PRTPRT

ROU

ROU

ROU

RUS

RUS

RUS

STP

SEN

SEN

SEN

SEN

SRB

SYC

SLE

SVN

ZAF

ZAF

ZAF

ESP

ESP

ESP ESP

KNA

SDN

TZA

THA

TGO

TGO

TTO

TUR

TUR

TUR

TUR

TUR

UGA

UKR

UKR

UKR

URY

URY

URY

URY

URY

URY

URY

VEN

VEN

VEN

VEN

VENVEN

VENVEN

VEN

VNM

YEM

SRB

SRB

SRB

ZMB

ZWE

0

50

100

1815 1835 1855 1875 1895 1915 1935 1955 1975 1995 2015

200 years of haircuts: 1815-2016 Meyer, Reinhart, and Trebesch (2019)

Circle size=debt volume (real 2009 US$)

Average Haircut: 44%(post-1970 average is 37%)Smaller than haircutfor US corporate debt.

Haircutsize

28

Mean: 8.4Median: 7Max: 30Obs: 95

Source: Clemens, Mayer, Reinhart, and Trebesch (forthcoming, 2020)

The first restructuring is oftennot enough to restore debtsustainability (≈ 70% require morethan 2 restructurings)

Mean: 2.3Median: 1Max: 8Obs: 83

In years

Debt restructurings with privatecreditors can drag on. For about2/3rds of cases it takes over 4 years.

Note: Cote D’Ivoire holds the post-WWII record (1983-2012)

Reinhart 29

Stylized crisis timeline of the 1920s/1930s and 1980s/1990sSource: Reinhart and Trebesch (2014)

Multiple restructurings were commonplace (Brazil had 6 in the 1980s and Poland 8)

Default, restructuring, and debt relief: World War I debt to the US and the UK, 1934, emerging markets, 1978-2010 (debt relief as a percent of GDP)Sources: Reinhart and Trebesch (2014 and 2016)

Reinhart 30

31

Real per capita GDP around debt relief events (exit from default) in middle income emerging markets (1978-2013) and advanced economies (1934)10-year window around debt relief event, level of real per capita GDP at T=1Sources: Reinhart and Trebesch (2014 and 2016)

Reinhart

0.80

0.90

1.00

1.10

1.20

1.30

T-5 T-4 T-3 T-2 T-1 T T+1 T+2 T+3 T+4 T+5

GDP for 15 advanced economy interwar episodes (T=1934)

GDP for 30 middle-to-high income emerging market episodes, 1978-2010 (T=year of final restructuring)

Index T=1

Reinhart

A note on banking crises—which are proliferating at the moment

Apart from dealing with sovereign debt challenges—many countries will be facing systemic banking crises.

Banking crisis resolution is importantly about dealing with private debt.

If I am asked the question: What factors made the 2008-2009 crisis so protracted in Europe?

My response includes: That an often neglected but key deterrent to recovery was the lack of deleveraging and write-downs of nonperforming assets even as late as almost a decade later.

32

Reinhart 33

Rapid deleveraging and banking crisis resolution in the aftermath of the Asian Crisis: External public and private

debt in 6 Asian economies, 1970-2013 (% of GDP)

15

25

35

45

55

65

75

85

1970 1975 1980 1985 1990 1995 2000 2005 2010

Banking crises in at least three

countries (shaded)

Average for India, Indonesia,

Korea, Malaysia, Philippines,

and Thailand

Percent

Sources: International Monetary Fund, World Economic Outlook, Reinhart and Rogoff (2009), Reinhart (2010), World Bank

(2013), International Debt Statistics, Washington DC http://data.worldbank.org/data-catalog/international-debt-statistics,

and World Bank, Quarterly External Debt Statistics, (QEDS),

http://web.worldbank.org/WBSITE/EXTERNAL/DATASTATISTICS/EXTDECQEDS/0,,menuPK:1805431~pagePK:64168

427~piPK:64168435~theSitePK:1805415,00.html

Faster recoveries from banking crises (relative to the Global Financial Crisis experience)were not limited to theAsian crisis episode.

Source: Reinhart and Tashiro (2014)

Some related reflections on crisis resolution(see also, How we handle bad debts will determine success of economic recovery, Reinhart (2020) https://www.hks.harvard.edu/faculty-research/policy-topics/development-economic-growth/dissecting-coronavirus-economic-impact)

Reinhart 34