Macroprudential Policy:Macroprudential Policy: What Instruments and How to Use Them?

Lessons from Country Experiences

Monetary and Capital Markets Department

Road MapRoad Map

C i i h d i lI. Country experiences with macroprudential instruments

II. Effectiveness of macroprudential instrumentsp

III Lessons and Policy MessagesIII. Lessons and Policy Messages

i iIV. Next Steps: Remaining Gaps

I. Country Experiences withI. Country Experiences with Macroprudential Instruments

What Instruments Are Used?What Instruments Are Used?

0 f h f l d i• 10 of the most frequently used instruments are examined

• Mostly prudential instrumentsy p

• Those capable of addressing systemic risk are• Those capable of addressing systemic risk are considered macroprudentialProcyclicality (time dimension)Procyclicality (time dimension)Interconnectedness (cross-sectional dimension)

The 10 InstrumentsThe 10 Instruments

• Credit related k d b d• Credit-related: – Caps on the loan-to-value (LTV) ratio – Caps on the debt-to-income (DTI)

ratio

• Risks generated by strong credit growth and asset price inflation;

ratio – Caps on foreign currency lending– Ceilings on credit or credit growth

• Liquidity-related:

• Systemic liquidity risk ;

Ri k i i f iLiquidity related: – Limits on net open currency

positions/currency mismatch (NOP)– Limits on maturity mismatch

• Risks arising from excessive leverage and the consequent de-leveraging; y

– Reserve requirements• Capital-related:

– Countercyclical/time-varying capital • Risks related to large and

volatile capital flows and y / y g prequirements

– Time-varying/dynamic provisioning– Restrictions on profit distribution

pcurrency fluctuations.

Objectives of Macroprudential Instruments

80

100

credit growth/ asset price inf lation

20

40

60

0 excessive leveragecapital f lows/ currency f luctuation

systemic liquidity risk

Caps on LTV Limits on maturity mismatch

Limit on net open currency positions/currency mismatch Restrictions on profit distribution

Source: IMF Financial Stability and Macroprudential Policy Survey, 2010y p y y

What Affects the Choice of Instruments?

(% of countries in each group using each type of instruments)(% of countries in each group using each type of instruments)

Credit-related Liquidity-related Capital-related

Advanced 43 19 10EconomicDevelopment

Stage

Advanced 43 19 10

Emerging Market

68 93 68g Market

Exchange Rate

Flexible 48 55 40

M d/Exchange Rate Managed/Fixed

100 89 56

f l Large 48 36 20Size of Financial Sector

Large 48 36 20

Small 67 88 67

llSize of Capital Inflow

Small 58 54 29

Large 56 68 56

How Are the Instruments Used?How Are the Instruments Used?

6436

MultipleSingle

4555

d

TargetedBroad-based

8

5545

Rule

Time-varyingFixed

4159

92

No coordinationCoordination

Discretion

41

0 20 40 60 80 100

No coordination

(percent)

II. Effectiveness of Macroprudential II. Effectiveness of Macroprudential Instruments

Effectiveness of the InstrumentsEffectiveness of the Instruments

• Dampening pro-cyclicality of– credit growth?g

– leverage?

• Limiting interconnectedness in exposures tog p– wholesale funding?

foreign sources of funding?– foreign sources of funding?

Three ApproachesThree Approaches

1. The case study

2. The simple approach

3 The panel regression3. The panel regression

1 Case Study1. Case Study

ll b d f• Small but diverse group of countries: – China, Colombia, Korea, New Zealand, Spain, the

United States and some Eastern European countries.

• Instruments seem to have achieved, to variousInstruments seem to have achieved, to various degrees, their intended objectives – Effectiveness does not depend on size of financialEffectiveness does not depend on size of financial

sector or exchange rate regime

2 Simple Approach2. Simple ApproachChange in Credit Growth After the Introduction of Instruments

1.0%LTV(y/y change)

0.5%

1.0%DTI(y/y change)

Change in Credit Growth After the Introduction of Instruments (average across countries)

0.0%

0.5%

t-2 t-1 t t+1 t+2 t+3 t+4 -1.0%

-0.5%

0.0%

0.5%

t-2 t-1 t t+1 t+2 t+3 t+4

-1.5%

-1.0%

-0.5%

Quarterly-3.0%

-2.5%

-2.0%

-1.5%

QuarterlyQuarterly

5.0%

6.0%

7.0%Reserve Requirements(y/y change)

0.5%

1.0%

Dynamic Provisioning(y/y change)

1.0%

2.0%

3.0%

4.0%

-1.0%

-0.5%

0.0%t-2 t-1 t t+1 t+2 t+3 t+4

-2.0%

-1.0%

0.0%t-2 t-1 t t+1 t+2 t+3 t+4

Quarterly -2.0%

-1.5%

Quarterly

2. Simple ApproachCredit Growth vs. GDP Growth

p pp

With and Without Caps on Loan-To-Value Ratios With and Without Caps on Debt-to-Income Ratios

3%

4%

5%

With and Without Caps on Loan-To-Value Ratios

No Caps on LTV (blue)

Caps on LTV (red) 3%

4%

5%

)

With and Without Caps on Debt to Income Ratios

No Caps on DTI (blue)

Caps on DTI (red)

-1%

0%

1%

2%

-10% -5% 0% 5% 10%

wth

(Per

cent

Qua

rterly

)

GDP Growth (Percent Quarterly)-1%

0%

1%

2%

-10% -5% 0% 5% 10%

owth

(Per

cent

Qua

rterly

GDP Growth (Percent Quarterly)

-5%

-4%

-3%

-2%

Cre

dit G

row

With and Without Reserve Requirement -5%

-4%

-3%

-2%

Cre

dit G

ro

With and Without Dynamic Provisioning-5%

2%

3%

4%

5%

y)

q

No Reserve Requirements (blue)

Reserve Requirements (red)

3%

4%

5%

y)

With and Without Dynamic Provisioning

No Dynamic Provisioning (blue)

Dynamic Provisioning (red)

-1%

0%

1%

2%

-10% -5% 0% 5% 10%

row

th (P

erce

nt Q

uarte

rl

GDP Growth (Percent Quarterly) -1%

0%

1%

2%

-10% -5% 0% 5% 10%

row

th (P

erce

nt Q

uarte

rly

GDP Growth (Percent Quarterly)

-5%

-4%

-3%

-2%

Cre

dit G

-5%

-4%

-3%

-2%

Cre

dit G

r

3. Panel RegressiongEstimates of effects

Caps on LTV 80% 34% (NS) N/A N/A

Caps on DTI 100% 100% N/A N/ACaps on DTI 100% 100% N/A N/A

Limits on Credit Growth 100%> 80% N/A N/A

Limits on NOP N/A N/A 15% NS/ /

Limits on Maturity Mismatch N/A N/A NS 5%

Reserve Requirements 92% 100% N/A N/A

Time-varying/DynamicProvisioning

100%> 100%> N/A N/A

Countercyclical/Time-varying NS 100%> N/A N/ACountercyclical/Time varyingCapital Requirements

NS 100%> N/A N/A

N/A: Not ApplicableNS: Non-significant

3 Panel Regression: Caveats3. Panel Regression: Caveats

R i ffi i t• Regression coefficients are:

– Averages of country performancesAverages of country performances

– Affected by small sample size

– Not an indication of equal effectiveness in all countries

• Country-specific circumstances important for effectiveness

• Use of instruments is new: limited number of observations

III L d P li MIII. Lessons and Policy Messages

Lessons and Policy MessagesLessons and Policy Messages

Instruments may be effective addressing risks generated by:– Credit growth/asset price inflation: credit-related instruments

– Systemic liquidity risk: liquidity-related instrumentsliquidity related instruments

– Excessive leverage: capital-related instrumentscapital-related instruments

– Capital flows: all three types of instrumentsall three types of instruments

Lessons and Policy MessagesLessons and Policy Messages

• Useful to adjust the instruments at different phases of the cycle

• Instruments that vary through the cycle have• Instruments that vary through the cycle have advantages and should be used when possible

• Well coordinated policy actions: necessaryWell coordinated policy actions: necessary condition for success

Lessons and Policy MessagesLessons and Policy MessagesMacroprudential Instruments

How to use

Macroprudential Instruments

Multiple Fixed Time-varying Rules DiscretionSingle

Do's and Don'ts

Broad-based TargetedCoordination

with other policies

Use when risk is well-defined from a single

source

Do not overdo the use of multiple

instruments or

Use if granular data are not

available and risks are

Be ready to adjust fine-

tuning; anticipate

Establish mechanisms to resolve conflict

and assign clear

Adjust parameters if needed with

changing

Design sound and transparent

principles governing the

Use when risk of inaction is high and risk management

Use when have deep structural changes and

rapidly evolving impose costs

that are too highgeneralized

pchannels for

evasion

Supplement with broader-

based

gaccountability

and governance arrangements

g gcircumstances

g gadjustment

gand supervision capacity is weak

Re-assess calibration periodically

p y grisks

Do not overdo discretion

based measures as

needed to limit the scope for circumvention

periodically

Avoid excessive complexity

Next Steps: Remaining GapsNext Steps: Remaining Gaps

• Deeper analysis of interconnectedness (cross-section dimension) )– Data availability is a constraining factor.

• Deeper understanding of design and• Deeper understanding of design and calibration of instruments

• Estimates of cost of implementation: distortions, unintended consequencesdistortions, unintended consequences

• Relationship between macroprudential and i d i l l imicroprudential regulation