The Chilean Experience with (De)Dollarization

1 C E N T R A L B A N K O F C H I L E 15th December 2006

The Chilean Experience with (De)Dollarization

Kevin Cowan

Prepared for the conference “Dollarization: Consequences and Policy Options”

Istanbul, 15th December 2006

C E N T R A L B A N K O F C H I L E 15th December 2006 2

Outline

1. Why Chile?

2. Brief overview of determinants of dollarization – framework for discussion.

3. The Chilean experience.

4. Policy implications.

3 C E N T R A L B A N K O F C H I L E OCTOBER 6 2006

I. Why Chile?

C E N T R A L B A N K O F C H I L E OCTOBER 6 2006 4

Why Chile?

After the early 80s, low levels of (domestic) financial dollarization, despite history of high inflation => relatively low currency mismatches and associated vulnerabilities:

Banking sector

Government sector

Corporate sector


Bank Dollarization

Source: Data for 2001. De Nicolo et al (2003), Arteta (2002) and Bank Superintendency of the Dominican Republic

Argentina

Chile Honduras

Haiti

Bolivia

Costa Rica

Guatemala

JamaicaMexico

Nicaragua

Peru

Paraguay

Uruguay

Venezuela

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100

Share of dollar deposits (in %)

Share o

f d

ollar loans (

in %

)

Turkey

Thailand

Indonesia


Dollarization of Public Debt

Source: CLYPS 2006

ArgentinaBoliv ia

Brazil

Chile

ColombiaCosta Rica

Honduras

Jamaica

Mexico Nicaragua

Peru

Uruguay

Venezuela0

.25

.5

.75

1

Share

of D

om

estic D

ebt in

US

D

0 .25 .5 .75 1Share of Deposits in USD

Bank and Government Dollarization


Corporate Dollarization

Source: IADB 2004

Firm Level Liability Dollarization in Tradable and Non-Tradable Sectors in Selected Latin

American Economies

(Median Value, Year 2001)

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Brazil

(0.0%)

Colombia

(0.59%)

Chile

(10.7%)

Mexico

(10.8%)

Costa Rica

(43.8%)

Peru

(66%)

Argentina

(73.6%)

Uruguay

(84.6%)

Tradable

Non-Tradable


Why Chile?

Deep capital markets by EME standards:

Private bank loans (70% of GDP)

Large (but iliquid) stock market (90% of GDP)

Relatively developed (but also iliquid) private bond market (20% of GDP or which13% corporate)

Note, however, that 100% of foreign issued debt is in “hard” currencies.

Low dollarization has not come at the expense of short tem debt either.


Maturity of Bonds

Source: Braun and Briones 2006 based on SDC data

GuaVenPol

CosEcu

Hun

ColHon

Sin

Slo

Bra

Ind

Tha

Arg

PerPhi

Cze

SweUS

GerPan

Mex

Bel

New

Por

Gre

Mal

Fra

Nor

Aus

Lie

NetFinIre

Swi

Chi

LuxBol

Ita

Aus

Den Can

Uru

Spa

CHL

0.2

.4.6

.81

Sh

are

Bo

nd

s O

ver

Fiv

e Y

ears

0 5 10 15Average Maturity

Bonds Issued 1995-2004

Maturity of Private Sector Bonds


Maturity of Corporate Debt

Source: IADB (2004)

ARG

BOL

BRA

CHL

COL

CRI

MEX

PERURY

.5.6

.7.8

Shar

e of

Short

Ter

m D

ebt

0 .2 .4 .6 .8Share of Dollar Debt

median 1994 - 01

Maturity and Currency Corporate Sector


II. Causes of Financial Dollarization


What drives dollarization?

To a large extent domestic financial dollarization can be explained by inflation uncertainty. If inflation risk is high, better to take on the real exchange rate risk of dollar contracts.

This is the central result of the work on Minimum Variance Portfolios (MVPs) by Ize and Levy-Yeyati (2003) and De Nicolo , Honohan and Ize (2004).


MVP and Dollarization (LA)

Source: IADB (2004)

SLV

HNDTTO

PRY

CHL

HTI

MEX

ECU

COL

CRI

JAM

GTM

VEN

ARGPER

NIC

BOL

URY

-40

-20

020

40

doll

ariz

atio

n o

f dep

osi

ts

-.4 -.2 0 .2 .4 .6mvp

coef = 65.927968, (robust) se = 10.589308, t = 6.23

Deposit Dollarization and MVP in LAC 1999

Note: Chile is an outlier (significantly so)



Therefore, part of dollarization is “warranted” (conditional on agents priors of the distribution of prices and RER).

Explains high correlation STD and dollar debt.

This suggests that: Policies that restrict dollarization, by

increasing the risk of saving/borrowing, will lead to lower intermediation.

Cowan & Kamil (2004) find that this is indeed the case (dif-dif approach)


Cost of Restricting Dollar Debt

Variance of

real returns of

portfolio

Share of

dollar debt

MVP

Minimum

variance

Variance of

Inflation

0

Variance loss

Restricting dollar debt has a larger effect on

size of domestic financial market in countries with large

“variance loss”



Regarding foreign debt empirical evidence (Eichengreen et al 2003) suggest that domestic macro outcomes have scarce impact => “original sin”.

Note, however: Recent issuance in “pesos” (Tovar 2005)

Growing importance deliverable forwards (Selaive et al 2006)

Despite low variance in EMEs as to debt composition, high variance in overall composition of gross foreign liabilities.


Reducing Dollarization

I. Restrictions => desintermediation

II. Building credibility => timely (and learning possibly hindered by dollarization itself)

III. Indexation: for moderate inflation levels, better hedge than dollar => many unsuccessful attempts (Galindo and Leiderman 2005), additional macro costs.


III. Dollarization in Chile


A Brief History of Indexation in Chile: The Unidad de Fomento

1959: Centralized Savings and Loans system established for low income housing. Savings indexed to CPI or wages, annual.

1960: Following financial deregulation SINAP and CCAP (private and public savings and loans agencies set up. The CCAP indexed loans to CPI, latter extended to SINAP.

64-70: Active promotion of indexation:

65: Large share of state bank deposits indexed.

64: Central bank issues indexed bonds (CAR)

UF created (later pefected to monthly and daily…)


Bank Dollarization

Source: CBCh

0%

10%

20%

30%

40%

50%

60%

70%

1978

1980

1982

1984

1986

1998

1990

1992

1994

1996

1998

2000

2002

US$

UF

Pesos

0%

10%

20%

30%

40%

50%

60%

70%

80%

19

77

19

79

19

81

19

83

19

85

19

87

19

89

19

91

19

93

19

95

19

97

19

99

20

01

20

03

US$

Pesos

UF

•Three periods: 1) Loan dollarization late 1970s

2) Extensive indexation 1983-2000

3) Growing nominalization 2001

Loans Deposits

NB: Deposit doll. low


Late 70s: Loan Dollarization

Late 70s, remaining quantitative restrictions on capital inflows by banks lifted.

Fixed exchange rate + currency blind bank regulation (necessarily so).

Implicit bailouts + extensive related lending.

Large private inflows intermediated by banking system => significant mismatches.


Currency Mismatches

Agriculture 0.20

Mining 0.36

Industry 0.43

Construction 0.29

Commerce 0.26

Transport and Telecom. 0.21

Financial Services 0.09

Share of Bank Loans to Each Sector in Dollars

Source: SBIF


The UF period

Financial crisis of 82 marks transition to next stage.

Large scale bailout of financial system, tied to UF:

UF bonds given to banks whose dollar loans were paid with the preferencial (low) dollar.

Domestic debt restructured to UF.

CBCh grants UF credit lines.


The UF period

Real macro policy supportive of indexation:

Real (UF) interest rate target,

Exchange rate band aimed at stabilizing UF/dollar exchange rate,

Gradual reduction of inflation,

Issuance of CBCh debt in UF => long + liquid yield curve for UF instruments.

Early 80s: pension reform => “captive” demand for UF indexed debt.


The UF period

Stable RER and high domestic rates => large capital inflows.

Reflected in part in an upward trend in corporate external borrowing in 90s

Pervasive real indexation, in particular for wages:

2 year contracts

6 month indexation


Nominalization

Transition again marked by period of instability (1998=>) following Asia Crisis.

Macro preconditions:

Fiscal prudence over almost 3 decades

Steady fall in inflation (partial targeting since early 90s)


Inflation in Chile

-5

0

5

10

15

20

25

30

ene/

90

ene/

91

ene/

92

ene/

93

ene/

94

ene/

95

ene/

96

ene/

97

ene/

98

ene/

99

ene/

00

ene/

01

ene/

02

ene/

03

ene/

04

ene/

05

ene/

06

(%)

Piso

Techo

IPC (var % en 12 meses)


Nominalization

Exchange rate band abandoned in Sep.1999.

Fully fledged IT adopted in 2001.

Nominalization of CBCh monetary policy:

Nominal target rate (TPM),

CBCh issues peso bonds in short end of curve.


Nominalization

Increased exchange rate variance after float.

Falling corporate mismatches:

Less borrowing

More derivative hedging

Better matching: similar results by Martinez and Werner for Mexico and by Kamil (2006) for sample of LA economies.


Mismatches in the Chilean Corporate Sector Fall

Dollar debt adjusted by forward position

(% total assets)

4%

6%

8%

10%

12%

1995 1996 1997 1998 1999 2000 2001 2002

dollar debt dollar debt adj. dollar debt - forw. (dollar debt - forw.) adj.

Source: SVS, BCdeCH)


Nominalization + Float

Increased exchange rate variance after float.

Falling corporate mismatches: Less borrowing

Better matching => similar results by Martinez and Werner for Mexico and by Kamil (2006) for 6 LA countries.

More derivative hedging

Growth of FOREX derivative markets => consistent with cross country results by Selaive et al 2006.


Derivative Forex Market Grows

Notional Derivative Positions and Exchange Rate Volatility

Source: CBCH

0

2

4

6

8

10

12

14

16

18

20

90 92 94 96 98 00 02 04 06

0

20,000

40,000

60,000

80,000

100,000

120,000

Annualized SDev of $/USD (GARCH 1,1)

USD mill


Derivative Market Grows

Turnover in Spot and Derivative FOREX Markets 1998-2004

(as % of nominal GDP)

Source: Ahumada et al (2006)

9.06.0 5.6

17.2

8.8 9.56.4

4.1 4.2

17.1

15.414.7

24.9

17.219.0

0.8 1.6

3.2 2.6 2.4 2.4 1.4 2.71.7

2.4

2.5 3.22.9 3.13.1

0.9 0.8

0

5

10

15

20

25

30

35

40

45

1998 2001 2004 1998 2001 2004 1998 2001 2004 1998 2001 2004 1998 2001 2004 2005

Spot Derivatives

Industrial

Economies

Emerging

Economies

Emerging

Economies wo

SGP, HK

LAC Chile


FOREX risk markets

Banks and Pension Funds play an key role in these markets:

Banks intermediate

Pension funds provide long positions

0

5

10

15

20

25

30

35

96 97 98 99 00 01 02 03 04 05 06

Límite Inv extranj

Activos en MX

Exposición Neta MX

Exposición Neta US$

Límite max expuestoForeign currency

Foreign assets

Financial

Internediaries

Pension Funds

Foreigners

Companies

Other

Counterparts in Banks Derivative Positions

Foreign Assets in Pension Funds


III. Policy Lessons


Policy Conclusions

1. Domestic indexation played an important role in limiting domestic financial dollarization:

UF important Indexed unit

By early 80s 20 years of history + clean track record

But also supported by: Real macro policy + public yield curve in UFs

Real indexation (causal?) => large costs when faced by real shocks (e.x wage rigidity following 98 ToT shock)

Important public sector role: All initial indexation state related

Key role played by UF in rescue of financial system

2. Context of prudent macro policies


Policy Conclusions

Domestic nominalization:

1. Nominalization of CBCh monetary policy and IT framework have coincided with falling share of short term UF debt contracts:

Credibility => inflation track record + IT have reduced inflation uncertainty in medium term horizon.

Peso yield curve (up to 10 years).

Nominal target rate (stabilizes nominal rates…).

2. Does not imply need to eliminate UF (works very well for long term contracts!)


Policy Conclusions

Foreign dollarization

1. In Chile less an issue of the currency composition of external debt than level and allocation of risk in the economy: Exchange rate regime has played important

role (pre 82, post 99): float has reduced borrowing and improved matching.

Bank regulation played key role in 1982 episode.

2. Current policy stance: attract foreigners to domestic market rather than issue abroad..

39 C E N T R A L B A N K O F C H I L E 15th December 2006

The Chilean Experience with (De)Dollarization

Kevin Cowan

Prepared for the conference “Dollarization: Consequences and Policy Options”

Istanbul, 15th December 2006

Documents

The Chilean Experience with (De)Dollarization