Surprising Similarities: Recent Monetary Regimes of Small Economies

Surprising Similarities:Recent Monetary Regimes

of Small Economies

Andrew K. RoseBerkeley-Haas, CEPR and NBER

Key Question

• Global Financial Crisis (GFC) spread globally, especially via capital (and trade) flows

• How did outcomes for small economies vary by monetary regime?– Did one monetary regime provide more insulation

from GFC spillovers?

Rose: Monetary Regime Similarities 2

Focus

1. Recent Period– Panel of data, mostly focusing on period of and

since GFC (2007-2012)

2. Small Countries– Who’s Small? The non-large (>170 countries)

• Exclude the IMF Systemic-5• China, EMU, Japan, UK, USA


Key Finding 1

• Monetary regimes of small economies have been stable recently

• Unexpected because:– Historical counter-cyclicality– Size of GFC and Great Recession

• Why the Stability? What’s New?– Countries that Float with Inflation Targeting– (Hard Fixes always around)


Key Finding 2

• Remarkably Similar Recent Outcomes for Hard Fixers and Inflation Targeters– Business Cycles, Inflation, Capital Flows,

Responses to Capital Flows …

• Implausible or Boring?– Surprising since these appear to be two very

different monetary regimes– Banal for literature (since Baxter-Stockman), since

regimes matter for little except exchange rate


Data Set

• Annual Panel• Standard sources (WDI, Polity IV, EFW, IFS,

BIS)


Measuring the Monetary Regime

1. Issue: de jure or de facto?2. Issue: exchange rate or monetary regime?3. Issue: want recent span• Solution: de facto monetary regimes from

IMF Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER)– Potential problem: short span (2001-2012)


de facto Alternatives

• Levy-Yeyati and Sturzenegger– Exchange rate regimes through 2004

• Reinhart and Rogoff– Exchange rate regimes through 2010

• Stone and Bhundia– Monetary Regimes 1990-2005


Monetary Regimes of the Small

• AREAER: country-years split into 44 cells– Depends on

• Degree of exchange rate rigidity• orientation of fix (if any)• objective of float (inflation, money aggregate, ‘other’)

• My focus: two extreme monetary regimes– Floating with Inflation Target

• AREAER criteria deliver similar results to Mishkin’s 5

– Hard Fix


2006 Monetary Regimes for Small

• 26 float with inflation target– 1 left by 2012 (Slovakia for EMU)

• 83 hard fix– No separate legal tender; currency board;

conventional peg• (Conv. Peg: Bahamas, Denmark, Latvia, Saudi Arabia,

Benin, Namibia …)• (Insensitive to inclusion: appendix)

– 23 left by 2012


Inflation Targeters Continuously 2006-12

Armenia Australia Brazil Canada Chile

Colombia Czech Republic Guatemala Hungary Iceland

Indonesia Israel Korea, Rep. Mexico New Zealand

Norway Peru Philippines Poland Romania

South Africa Sweden Switzerland Thailand Turkey


Inflation Targeter in 2006,Exited by 2012


Slovakia (Joined Euro)

Hard FixersContinuously 2006-12


Antigua & Barb. Aruba Bahamas Bahrain Barbados

Belize Benin Bhutan Bosnia & Herz. Brunei

Bulgaria Burkina Faso Cameroon Cape Verde Cen. Afr. Rep.

Chad Comoros Congo, Rep. Cote d'Ivoire Denmark

Djibouti Dominica Ecuador El Salvador Eq. Guinea

Eritrea Fiji Gabon Grenada Guinea-Bissau

Hong Kong Jordan Kiribati Latvia Lesotho

Libya Lithuania Mali Marshall Islands Micronesia

Montenegro Morocco Namibia Nepal Niger

Oman Palau Panama Qatar Samoa

San Marino Saudi Arabia Senegal St. Kitts & Nevis St. Lucia

St. Vinc. &Gren. Swaziland Timor-Leste Togo Unit Arab Emir.

Hard Fixers in 2006,Exited by 2012


Azerbaijan (Float, Other) Mauritania (Float, Other)

Belarus (Float, Other) Pakistan (Float, Mon Targ)

Cyprus (Euro) Seychelles (Float, Mon Targ)

Egypt (Float, Other) Slovenia (Euro)

Estonia (Euro) Solomon Islands (Float, Other)

Honduras (Soft Fix) Syria (Soft Fix)

Lebanon (Soft Fix) Trinidad & Tobago (Soft Fix)

Macedonia (Soft Fix) Ukraine (Float, Mon Targ)

Maldives (Soft Fix) Vanuatu (Soft Fix)

Malta (Euro) Vietnam (Soft Fix)

Hard Fixers in 2006 and 2012but not continuously in between


Kuwait Turkmenistan Venezuela

“Sloppy Center” Remaining

• Soft fixes (stabilized, crawling peg/band, pegged exchange rate within horizontal bounds);

• Floating with a monetary target (variants including crawl-like, managed, or free floats)

• Floating ‘Other’• 32 in 2006-2012 (17 with AREAER switches)

– 30 switched out by 2012


Sloppy CenterContinuously 2006-12

(* indicates switched IMF de facto Monetary regime)


Afghanistan Algeria* Botswana Burundi* Cambodia*

Congo, D. Rep.* Costa Rica* Gambia, The Guinea* Haiti*

India Iraq* Jamaica* Kenya Kyrgyzstan*

Lao PDR* Liberia* Madagascar Malaysia Mauritius*

Mozambique Myanmar* Nicaragua Pap. N Guinea* Paraguay*

Singapore* Somalia Sudan* Tanzania Tonga

Uganda Zambia

Sloppy Center in 2006, Exited by 2012


Albania (Inflat. Target) S Tome & Prin (Hard Fix)

Dom Repub. (Inflat. Target) Serbia (Inflat. Target)

Georgia (Inflat. Target) Uruguay (Inflat. Target)

Ghana (Inflat. Target) Zimbabwe (Hard Fix)

Moldova (Inflat. Target)

Sloppy Center in 2006 and 2012 but not continuously in between


Angola Argentina Bangladesh Bolivia Croatia

Ethiopia Guyana Iran Kazakhstan Malawi

Mongolia Nigeria Russia Rwanda Sierra Leone

Sri Lanka Suriname Tajikistan Tunisia Uzbekistan

Yemen

Durability of Monetary RegimesSmall Economies

--- Monetary Regime in 2006 ---

2012 Mon. Regime Inflat. Targeting (26) Hard Fix (83) Sloppy Center (62)

Inflation Targeting 25 0 7

Hard Fix 0 63 2

Sloppy Center 0 16 53

EMU Entrants 1 4 0

% continuously in regime since 2006

96% 72% 23%

% 2011 Global GDP 20% 4% 7%


Sloppy Center: Turnover is Big and (Counter-) Cyclic

# Shifts in IMF De FactoMonetary Regime

Global RealGDP Growth

2002 27 2.92003 8 3.72004 7 5.02005 9 4.62006 7 5.32007 11 5.42008 28 2.82009 37 -0.62010 7 5.22011 21 4.02012 8 3.2


Representative?

• Since 1970, regime turnover has been high and (counter-) cyclic


Durability of Monetary Regimes , Small Economies: Historical Volatility

(excludes Germany, Japan, UK, USA)

% Continuously in same: All Countries Fixers

Reinhart-Rogoff regime, 1970-76

55% 59%

(% 1976 Global GDP) 12.3% 6.4%

Reinhart-Rogoff regime,1980-86

60 % 75%

(% 1986 Global GDP) 28.4% 3.3%

Levy-Yeyati-Sturzenegger regime, 1980-86

53 % 58 %

(% 1986 Global GDP) 9.1% 9.1%


Historical Cyclicality of Monetary/Exchange Rate Regimes

Business Cycle Measure:

GDP:HP

GDP:BK

GDP:CF

GDP: Growth

GDP: Linear

Unemp.Rate

Stone-Bhundia (1990-2005)

-14.9**(4.0)

-16.7**(4.2)

-12.9**(4.0)

-.04(.02)

-5.3**(1.7)

.13*(.05)

Levy-Yey.-Sturzen. (1974-2004)

-1.6(1.2)

-2.0(1.2)

-2.1(1.2)

-.01(.01)

-1.4**(.4)

.04(.03)

Reinhart- Rogoff (1970-2006)

-1.7(1.1)

-1.9(1.1)

-2.0(1.1)

-.003(.007)

-1.1**(.4)

.07*(.03)


Conclude: Stability from Inflation Targeting Regimes

• New• Significant• Acyclic


Monetary Regimes by the Numbers: Counting Countries


Inf. Target Float

Sloppy Center

Hard FixersGlobal Growth

-20

24

6

05

01

001

50N

umb

er o

f Eco

nom

ies

2001 2007 2012

Glo

bal G

row

th

Large Economies excluded: China, EMU, Japan, UK, and USGrouping Small Economies by Monetary Regime

Monetary Regimes by the Numbers: Sizing Up the Economies


Inf. Target Float

Sloppy Center

Hard Fixers

Global Growth

-20

24

6

0.2

.4F

ract

ion

of G

loba

l GD

P

2001 2007 2012

Glo

bal G

row

th

Large Economies excluded: China, EMU, Japan, UK, and USGlobal GDP of Small Economies by Monetary Regime

Tangent: Determinants of Monetary Regimes

• Remarkably little known– Recent surveys in Klein and Shambaugh (2010),

Rose (2011)

• Literature: focus on (sluggish) levels• Here: use Stone-Bhundia regimes (through

2005, pre-dates GFC)– Size, income, institutions, openness


Characteristics Across Monetary Regimes: Univariate Evidence

Averages: Sloppy Center Hard Fix Inflation Target Hard Fix = Inflation

Target (t-test)Log Population 16.5 15.5 16.6 -8.3**

Log R GDP p/c 8.9 9.4 9.7 -4.6**

Polity 5.7 3.1 9.4 -10.1**

Trade %GDP 84.5 88.8 68.7 5.0**

Chinn-Ito .4 .7 1.3 -4.3**

Inv Free, EFW 59.8 57.4 66.4 -5.2**

Fin Free, EFW 52.9 57.5 66.5 -4.8**

M2 %GDP 52.6 59.9 67.3 -2.1*

Credit %GDP 198.8 186.5 90.9 1.0


Quantile Plots

• Compare (univariate) distribution of key characteristics across regimes

• Quantiles for different regimes plotted on different axes

• Similar distributions: data plotted along diagonal



67

89

1011

Har

d F

ixer

s

7 8 9 10 11Inflation Targeters

Log Real GDP p/c

1012

1416

1820

Har

d F

ixer

s


Log Population-1

0-5

05

10H

ard

Fix

ers

5 6 7 8 9 10Inflation Targeters

Polity

010

020

030

040

0H

ard

Fix

ers


Trade, %GDP

Quantile Plots for Small Economies, 2011

Determinants of Monetary Regimes : Multivariate Evidence

Pop=0 .00** .00** .00** .00** .00** .00**

Income=0 .06 .03* .14** .29 .05* .03*

Polity=0 .00** .00** .00** .00** .00** .00**

Trade=0 .24 .27 .22 .38 .25

Fin Open=0

.56 .37 .28 .75 .20

Open=0 .37 .34 .34 .59 .15

Fix=IT .00** .02* .00** .00** .00** .01*

Obs 1108 1043 713 713 998 1074

Pseudo R2 .18 .20 .25 .25 .25 .19


Hard to extend Univariate Success

• Multinomial Logit compares determinants of three regimes


Determinants of Monetary Regimes : Coefficient Estimates

Hard Fix (versus Sloppy Center)

Log Population

-.33*(.14)

-.50**(.16)

-.58**(.18)

-.53**(.19)

-.60**(.17)

-.48**(.16)

Log Real GDP p/c

.52(.34)

.68*(.32)

.33(.36)

.33(.36)

.30(.32)

.59(.32)

Polity -.07*(.04)

-.09**(.03)

-.07(.04)

-.09*(.04)

-.10**(.03)

-.09**(.03)

Inflation Target Float (versus Sloppy Center)

Log Population

.60*(.26)

.50(.33)

.92**(.33)

.87**(.30)

.77*(.36)

.51(.33)

Log Real GDP p/c

.86*(.37)

.79*(.35)

.75*(.38)

.63(.41)

.96*(.39)

.95**(.37)

Polity .67**(.22)

.63**(.22)

.98**(.29)

.95**(.27)

.88**(.30)

.66**(.22)


Limited Success, No Surprises

• Small countries: more likely to (hard-) fix, less likely to target inflation

• Democracies: less likely to fix, more likely to target inflation

• Marginal effects of income• No effect of real or financial openness• Fit consistently poor!

– Always hard to determine monetary regime


Consequences of Monetary Regime

• Initially take regime choice as exogenous/random– Not so bad in practice– Will verify with IV/matching: results insensitive


Compare Three Regimes

• Inflation Targeting Floats– Ex: Brazil, Korea, Mexico, Canada

• Hard Fixes– 60 durable (Saudi Arabia, Hong Kong, Denmark)

• Sloppy Center– India, Russia, Iran


Compare long-lived, durable regimes (IT/HF)

1. With sloppy center2. With each other• Throughout use period since 2006


Quantile Plots: Outcomes for Hard Fixers Similar to Inflation Targeters• Unexpected a priori: two dis-similar monetary

regimes• But univariate distributions surprisingly alike,

much overlap



-50

050

Har

d F

ixer

s

-40 -20 0 20Inflation Targeters

Demeaned Real GDP Growth

-10

010

2030

Har

d F

ixer

s


CPI Inflation-5

00

50H

ard

Fix

ers

-30 -20 -10 0 10 20Inflation Targeters

Timor-Leste dropped

Current Account, %GDP

-20

-10

010

2030

Har

d F

ixer

s

-10 0 10 20Inflation Targeters

Government Budget, %GDP

Quantile Plots for Small Economies, 2007-2012


-20

020

4060

80H

ard

Fix

ers

-20 0 20 40 60Inflation Targeters

Capital Inflows, %GDP

-80

-60

-40

-20

020

Har

d F

ixer

s

-40 -20 0 20Inflation Targeters

Capital Outflows, %GDP-3

0-2

0-1

00

1020

Har

d F

ixer

s

-20 -10 0 10 20Inflation Targeters

Net Capital Flows, %GDP

-20

24

6H

ard

Fix

ers

-.5 0 .5 1 1.5Inflation Targeters

International Reserve Growth



-40

-20

020

40H

ard

Fix

ers

-20 -10 0 10 20Inflation Targeters

Real Eff Exch Rate Change

05

1015

20H

ard

Fix

ers

0 5 10 15Inflation Targeters

Long Bond Yield-1

000

100

200

Har

d F

ixer

s


Stock Market Change

-50

050

100

Har

d F

ixer

s


Property Price Change


Ditto Statistical Analysis

• Some differences between hard fixers and inflation targeters … but more similarities

• Not power: effects of HF/IT on inflation similar, significantly different from SC


Technical Stuff

• Panel LS on dummies for HF/IT– Omitted: sloppy center– Fixed time, random country effects (can’t do fixed)

• Regressands– Output consequences– Capital flows– Adjustment mechanisms

• Current account, capital controls, reserves, money, fiscal policy, real exchange rate appreciation, inflation


Effects of Monetary Regimes 2007-12: Regression Evidence on Output

Regressand InflationTargeting

HardFix

IT = H Fix?(P-value)

IT = H Fix = 0?(P-value)

BK-Filtered GDP .006(.004)

-.003(.004)

.04* .10

HP-Filtered GDP -.002(.001)

-.004*(.001)

.13 .04*

CF-Filtered GDP -.02(.02)

-.00(.04)

.77 .76

Demeaned Growth

-1.9*(.8)

-1.4(.8)

.56 .04*

Time-Detrended GDP

-.04(.03)

-.08**(.02)

.16 .01**


Effects of Monetary Regimes 2007-12: Regression Evidence on Capital Flows


HardFix



Gross Capital Inflows

3.2(3.2)

-4.1(6.4)

.90 .57

Gross Capital Outflows

-.0(3.2)

-3.2(6.7)

.61 .87

Net Capital Flows

3.2(1.9)

.8(1.6)

.03* .09

Std Dev Capital Inflows (c/s)

5.5(4.2)

5.5(6.9)

1.0 .38

Std Dev Capital Outflows (c/s)

5.1(4.2)

7.0(7.4)

.82 .36


Effects of Monetary Regimes 2007-12: Regression Evidence, Adjustment 1


HardFix



Current Account

1.6(1.4)

3.4(5.5)

.73 .49

ExportGrowth

.01(.01)

.00(.01)

.70 .85

ImportGrowth

-.00(.01)

.00(.01)

.76 .94

Chinn-Ito Capital Mobility

-.1(.4)

-.5(.3)

.41 .24

Fin Freedom Change

.01(.01)

.00(.01)

.16 .16

Inv Freedom Change

.03**(.01)

.01(.01)

.01** .01**


Effects of Monetary Regimes 2007-12: Regression Evidence, Adjustment 2


HardFix



M2 Growth (%GDP)

-.01(.01)

.00(.01)

.18 .41

International Reserve Growth

-.4(.4)

-.5(.4)

.26 .44

Government Budget

.3(.8)

.7(.9)

.70 .74

Change in Budget

.5(.7)

-.4(.5)

.30 .57


Effects of Monetary Regimes 2007-12: Regression Evidence, Prices


HardFix



CPI Inflation -4.4**(.7)

-5.2**(.6)

.15 .00**

GDP Inflation -4.7**(.8)

-5.2**(.7)

.41 .00**

Real Effective Exchange Rate

-15.0(9.8)

-20.1*(9.6)

.13 .05*

Chg Real Effect Exchange Rate

-3.9(3.4)

-5.4(3.5)

.06 .07

Growth in Stock Prices

-4.5(3.5)

-11.8**(3.3)

.01** .00**

Bond Yields -1.0(.8)

-1.0(1.0)

.96 .43

Growth in Property Prices

2.3(4.8)

-1.1(5.1)

.35 .63Rose: Monetary Regime Similarities 49

Sensitive to Estimation Technique?

• Use IV (size and polity: determination model)– Business cycle, investment freedom results

disappear– Similar results for Inflation (*), stock prices

• Also use nearest-neighbor matching– Match on government spending (% GDP),

unemployment rate– Similar results, especially for inflation


Summary

• Strong, Consistent Findings– Inflation lower (5%) for both hard fixers and

inflation targeters (than SC)– Net (but not gross) capital flows, change in

investment freedom higher for IT than HF

• Weak, Inconsistent/Marginal Finding– Worse business cycles, more real depreciation for

both HF and IT (than SC) – Stock prices declined more for HF than IT


Most Intriguing Result

• Rarely significant difference between effects of inflation targeting and hard fixes

• Plausible? – These monetary regimes seem to vary wildly

• Amount of discretion• Role of exchange rate


Consistent with Literature

• Rey (2013, italics added):“… analyses suggest monetary conditions are transmitted

from the main financial centre to the rest of the world through gross credit flows and leverage, irrespective of the exchange rate regime. This puts the traditional “trilemma” view of the open economy into question. Fluctuating exchange rates cannot insulate economies from the global financial cycle, when capital is mobile.”

– Note: Rey does not actually test relevance of exchange rate regime


Monetary Regime Matters Surprisingly Little: An Old Idea

• Klein and Shambaugh (2010), and references stretching back to …

• Baxter and Stockman (1989):– “Aside from greater variability of real exchange

rates under flexible than under pegged nominal exchange-rate systems, we find little evidence of systematic differences in the behavior of macroeconomic aggregates or international trade flows under alternative exchange-rate systems.”


Summary

• Small economies that float with an inflation target behave similarly to hard fixers over the post-bubble period – Artifact of methodology? (used a few)– Small data set? (>160 countries, 6 volatile years)– Literature unable to find significant differences

across monetary regimes• Perhaps there simply are few


Conclusion: 1

• Bulgaria and Romania:– Small open EMs in EU, reasonable improving

institutions, similar income and size– Bulgaria in hard fix; Romania floats with IT

• Many other examples– Denmark and Sweden; Ecuador and Colombia; El

Salvador and Guatemala; Cote d’Ivoire and Ghana; Bosnia-Herzegovina and Albania

• Similar countries choose different monetary regimes!


Conclusion: 2

• Differences in monetary regimes can persist because they’re unimportant– Growth, output gap, inflation, fiscal policy, current

account, reserve growth … do not vary much with monetary regime

– So “insulation value” of different monetary regimes similar in practice

– Consistent with literature if initially counter-intuitive (boring more than implausible)


Conclusion: 3

• Small countries now have new stable alternative to hard fix; float with inflation target– Both hard fixers and inflation-targeting floats have

survived GFC and aftermath• This monetary stability new (compare: 1930s/1970s)• Since countries in different monetary regimes behave

similarly, can’t easily differentiate between regimes


Documents

Surprising Similarities: Recent Monetary Regimes of Small Economies