Exchange Rate Regimes
Andrew K. Rose
Visiting Norman-Houblon Fellow
August, 2010
1
Three Questions
• Past: What do We Know about Exchange Rate
Regimes Historically?
– Rose “Fixed, Floating and Flaky”
• Present: How did Different Exchange Rate
Regimes do in the “Great Recession”?
– Rose and Spiegel “Causes and Consequences”
• Future: What are the Expected Effects of
China’s Switch in Regime?
– Eichengreen and Rose “27 Up”
2
Past: A Stylized Description of
Exchange Rate Regime
3
Exchange Rate Classifications
• Bad Old Days: IMF used official policy
• But De Jure Systems of Exchange Rate
Classification do not Cohere well with Actual
De Facto Behavior
4
3 Popular (Newish) De Facto
Classifications
• Levy-Yeyati and Sturzenegger
– Cluster Analysis on Exchange Rates and Reserves
• Reinhart and Rogoff
– Black Market Rates
• Shambaugh
– Nominal exchange rate movements
5
Poor Coherence
IMF
IMF
Levy-Yeyati &
Sturzenegger
Reinhart &
Rogoff
Shambaugh
Levy-Yeyati &
Sturzenegger
Reinhart &
Rogoff
100%
59%
100%
59%
55%
100%
68%
65%
65%
Shambaugh
100%
6
Message
• Don’t rely on any single system!
7
Some Stylized Facts
8
Many Countries are Fixed
Exchange Rate Regimes over Time
Distribution of Countries by Regime
Float
Float
Intermediate
50%
Intermediate
50%
Levy-Yeyati & Sturzenegger
100%
100%
IMF De Jure
Fix
0
0
Fix
1970
1980
1990
2000
1970
1980
1990
2000
2010
Reinhart & Rogoff
100%
100%
Shambaugh
2010
Float
Non-Peg
50%
50%
Intermediate
Peg
0
0
Fix
1970
1980
1990
2000
2010
1970
1980
1990
2000
2010
9
Not Much GDP in Fixers
Exchange Rate Regimes over Time
Distribution of GDP by Currency Regime
Levy-Yeyati & Sturzenegger
100%
100%
IMF De Jure
50%
Float
50%
Float
Intermediate
Intermediate
Fix
1970
1980
1990
0
0
Fix
2000
1970
1980
1990
2000
2010
Reinhart & Rogoff
100%
100%
Shambaugh
2010
Float
50%
50%
Non-Peg
1970
1980
1990
Fix
0
0
Peg
Intermediate
2000
2010
1970
1980
1990
2000
2010
10
Regimes are Becoming Durable
Exchange Rate Regime Switches over Time
Proportion of Global GDP in Economies with Changing Regimes
.4
.2
0
0
.2
.4
.6
Levy-Yeyati & Sturzenegger
.6
IMF De Jure
1970
1980
1990
2000
2010
1970
1990
2000
2010
.4
.2
0
0
.2
.4
.6
Reinhart & Rogoff
.6
Shambaugh
1980
1970
1980
1990
2000
2010
1970
1980
1990
2000
2010
11
Does Size (Population) Matter?
• Many fixers are small.
• But: many small economies
– Berkeley California has population > 49 (/237 ) “countries and
other entities” in CIA’s World Factbook
– Many included in the various exchange rate classifications
• No doubt that smallest economies of the world do not float
– Many don’t have own currencies
– 95 of CIA’s listed “countries” do not have national currency
• Easy to overstate; countries do not have to be large before
creating a floating currency
– Small floats include: the Seychelles (population 88,000 in June
2010), Tonga (123,000), and Sao Tome and Principe (176,000).
12
Quantile Plots of Size
Size of Fixers and non-Fixers
Quantile Plots of logs 2004 PWT 6.3 Population
Levy-Yeyati & Sturzenegger
4
9
Fix
9
4
Fix
14
14
Official IMF
4
9
non-Fix
14
4
14
4
9
Fix
9
4
Fix
14
Reinhart & Rogoff
14
Shambaugh
9
non-Fix
4
9
non-Fix
14
4
9
non-Fix
14
13
So Size Matters only at the tail
• Size Matters much less at 2.5 million
– 135 countries
• Size Matters not at all beyond 10 million
– 75 countries
14
Income? No Effect
Income of Fixers and non-Fixers
Quantile Plots of log 2004 PWT 6.3 Real GDP per capita
6
7
8
9
non-Fix
10
6 7 8 9
Fix
6 7 8 9
Fix
10 11
Levy-Yeyati & Sturzenegger
10 11
Official IMF
11
7
10
11
10 11
6
7
8
9
non-Fix
10
11
6 7 8 9
Fix
6 7 8 9
Fix
9
non-Fix
Reinhart & Rogoff
10 11
Shambaugh
8
6
7
8
9
non-Fix
10
11
15
Other Stylized Facts, 1
1. Two Anchors
•
Dollar (66 fixers); Euro (227)
2. All Large Rich Economies Float
•
•
•
•
Large: Dollar, Yen, Euro,
Medium: UK, Canada, Australia, Switzerland, …
EMs: Brazil, India, Indonesia, Korea, Mexico, Russia, and
Turkey
China is exception
16
Other Stylized Facts, 2
3. Regions Differ
– Sub-Saharan Africans fix
– Central Europeans, Asians do not
4. Oil Exporters Fix
– Especially OPEC members in Gulf
17
Other Stylized Facts, 3
5. Small Financial Centers Fix
– Mostly Small
6. Inflation Targeters Float
– Often Very Cleanly
7. Nominal Exchange Rate Volatility is Real
– Mussa
18
Causes of Exchange Rate Regime
• Theory #1: “Sources of Shocks”
– Countries with real shocks should float
– Financial shocks implies fix
• Stockman (2000) “the evidence supporting the
predictions of these models is only slightly better
than the evidence for cold nuclear fusion”
19
Another Theory: Credibility
• Fixed nominal exchange rate transparent easily
monitored monetary anchor
– Import credibility by fixing to Fed/Buba
• Tornell and Velasco: fiscal indiscipline eventually
undermines most fixes
– Float: easier to monitor, faster punishment, better
discipline
• So credibility arguments theoretically ambiguous
– Is exchange rate constraint different from other
constraints (e.g., Inflation Targeting)?
20
Microeconomic Arguments
• Facilitate Trade
– Size? Possible if Hedging Risk difficult (LDCs)
• Deepen Micro-Structure of FX market?
– Deepen liquidity
• But many rich countries (Denmark, HK) fix
• Little intervention outside FX (stocks, bonds)
21
Shameful Empirics
• No Time-Series Understanding
– OK since most determinants sluggish
• No Cross-Country Success Either
– Very small countries, autocracies , former
colonies, financial centers, oil exporters fix
– Little of the cross-country variation explained
though
• An Embarrassment! Almost no covariates of
exchange rate regime choices empirically.
22
What about Consequences?
23
Growth Consequences of Regimes?
Classification
Official IMF
Reinhart and
Rogoff
Levy-Yeyati
and
Sturzenegger
Narrow Crawl
.8*
(.3)
-.3
(.4)
Wide Crawl
.5
(.4)
-1.0*
(.5)
Intermediate
Float
-1.5**
(.4)
-.5
(.4)
Float
.2
(.5)
.5
(1.2)
Falling
-4.3**
(.6)
Non-Peg
Shambaugh
.3
(.3)
24
Inflation Consequences of Regimes?
Classification
Official IMF
Reinhart and
Rogoff
Levy-Yeyati
and
Sturzenegger
Shambaugh
Narrow Crawl
-9.1**
(2.1)
.4
(2.4)
Wide Crawl
2.7
(3.6)
.8
(3.1)
Intermediate
Float
18.4**
(3.1)
3.5
(1.9)
Float
8.8
(6.3)
7.9
(4.3)
Falling
62.**
(9.6)
Non-Peg
7.3**
(1.8)
25
Quick Summary of Consequences
• No Real Growth Effects
– Reasonable; monetary neutrality
• Unclear Inflationary Consequences
• Also: no effect on volatility
– Baxter-Stockman (1989); Flood-Rose (1995);
Obstfled-Rogoff “Six Puzzles” (2001)
26
Is this Question Worth Asking?
• Countries with similar income, size, openness,
institutions choose different regimes:
– Singapore vs. Hong Kong
– Denmark vs. Sweden vs. Finland
– Costa Rica vs. Panama
• No convergence, few apparent causes, no clear
consequences
27
Exchange Rate Regimes are Flaky
• Caring about exchange rate regimes is akin to
caring about individual preferences for wine
or beer
28
Present: The Non-Effect of
Exchange Rate Regimes on Crisis
Incidence
29
Cross-Country Approach
to “Great Recession
• Use cross-section of (107) countries to ask
which countries experienced biggest crises
• A necessary (but far from sufficient) part of
any successful early warning system
– Cross-sectional questions easier than time-series
modeling
• Attempt to link (2006 and earlier) crisis causes
to (2008 and later) crisis consequences
30
Rose-Spiegel (2010a,b,c) Findings
• ‘Great Recession’ progressive; countries with
higher income suffer worse crises (as in RS)
• No other robust results
– Over 80 “national” causes/vulnerabilities
– Over 40 “international” linkages
31
Cross-Sectional Data Set
• All countries/territories with real GDP per
capita at least $10,000 in 2003
• All countries/territories with real GDP per
capita at least $4,000 in 2003 and population
at least 1 million
32
Emerging Literature
• Four Big Differences from RS
– Measures of Crisis Intensity
– Potential Causes (Covariates)
– Estimator linking causes, intensity
– Country Sample
33
Key References
• Berkmen, Gelos, Rennhack, Walsh (2009)
“BGRW”
• Blanchard, Faruqee and Das (2010) “BFD”
• Claessens, Dell’Arriccia, Igan, Laeven (2010)
“CDIL”
• Frankel and Saravelos (2010) “FS”
• Giannone, Lenza and Reichlin (2010) “GLR”
• Lane and Milesi-Fettetti (2010) “LMF”
34
Measures of Crisis Intensity
(Dependent Variables)
1. Default: Real 2008-09 GDP growth (from EIU)
2. Real GDP growth change, 2008-09 - 2005-07
(LMF)
3. Real GDP growth change, 2008-09 - 1990-07
(BFD)
4. Revision to WEO 2009 growth forecast (BGRW)
5. 2009 output gap (from OECD)
6. 2008-09 consumption growth
7. First Principal Factor from 4 RS variables
35
Modeling Crisis Causes (Regressors):
Many Unsuccessful Attempts
• Credit %GDP
• Debt %GDP
• Domestic Banking Sector Characteristics
• Fiscal Policy
• Trade Flows
• Capital Flows
36
Causes: Some Successes
(RS Investigate All)
•
•
•
•
•
•
•
•
•
Fixed Exchange Rate Regime (BFD, LMF)
House Price Appreciation (CDIL)
Credit Growth (BGRW, CDIL, LMF)
Credit Market Regulation (GLR)
Current Account %GDP (CDIL, LMF)
Financial Leverage (BGRW)
Reserves (FS, Obstfeld et al for depreciation)
Short-Term External Debt (BGRW)
Trading Partner Growth (BFD, LMF)
37
Different Country Samples
•
•
•
•
•
•
Full sample (107)
(51) World Bank High Income
(74) IMF non-Advanced
(89) non-Oil Exporters
(91) non-Small Financial Centers
(51) non-oil, non-FC High/Upper-Middle
Income
38
Simple Econometrics
• LS cross-country regressions, White se’s
• Condition on log (2006) real GDP per capita
• Add dummy for 2006 fixers
• Different regressands, samples
39
Effect of Exchange Rate Regime
(Dummy for 2006 Fix)
All
2008-09
Growth
2008-09 Grow
- ’05-’07 Grow
2008-09 Grow
- ’90-’07 Grow
Revised WEO
‘09 Grow Fost
Output Gap
2009
Cons. Growth
2008-09
Extracted
Prin. Factor
-1.42
(1.41)
-2.11
(1.46)
-2.08
(1.42)
.16
(.72)
-.73
(.71)
.36
(2.39)
-.14
(.17)
High
Income
.51
(1.92)
-.25
(1.72)
-.60
(1.81)
1.15
(.98)
-.62
(.70)
3.46
(4.64)
.01
(.21)
No
Adv.
-1.79
(1.91)
-2.50
(1.99)
-2.48
(1.95)
-.13
(.91)
n/a
-1.06
(2.73)
-.17
(.24)
No
Oil
-3.07*
(1.48)
-3.21
(1.62)
-2.96
(1.53)
-.39
(.80)
-.73
(.71)
-1.47
(1.13)
-.26
(.18)
No Fin’l
Centers
-1.62
(1.55)
-2.42
(1.60)
-2.38
(1.56)
-.29
(.73)
-.55
(.78)
.38
(2.62)
-.22
(.17)
No Poor,
Oil, FCs
-2.29
(2.01)
-3.54
(2.24)
-2.76
(2.10)
-1.10
(.95)
-.55
(.78)
-1.86
(1.39)
-.21
(.22)
40
Note Insignificant Effects!
• Fixed Exchange Rate Regimes almost never
significantly affect growth
– Signs mostly negative (5/41 positive)
• Very weak indications of importance
41
True of Other Covariates Too!
• Very difficult to link cross-country crisis
incidence of causes and consequences of
“Great Recession”
42
Future: What Can We Expect
from China’s Switch?
43
Chinese Exit from Peg: June 19 2010
• What can We Expect?
• Are there any Precedents?
– Most departures from fixes occur under periods of
duress
– Accordingly, depreciation expected
– Modeled by Krugman and others
44
27 Precedents
• Use Reinhart-Rogoff monthly system of 15
exchange regime classifications
• China: pegged initially
– Moves to More Flexible Regime
– Appreciates
• 27 other observations of regime switch (fix to
more flexible) and appreciation ($ or SDR)
45
The 27 Episodes
Austria, 1971
Germany, 1971
Mauritania, 1974
S.Africa, 1972
Canada, 1970
Germany, 1973
Mozamb., 2004
Spain, 1974
Denmark, 1971 Hong Kong, 1972 Netherlands, 1971
Sweden, 1973
Eq. Guinea, 1979
Japan, 1973
NZ, 1973
Switzerland, 1973
Finland,1973
Libya, 1971
Nigeria, 1971
Tunisia, 1974
France, 1971
Malaysia, 2005
Portugal, 1973
UK, 1972
Germany, 1969
Malta, 1972
Singapore, 1973
46
Growth around Appreciating Exits
-20
-10
0
10
20
30
Annual Growth around (27) Fix Exits with Appreciation
-5
0
5
47
Little Happens to Growth
• Some heterogeneity
• Still, no reason to expect big changes
• What about the macro-economy in general?
48
The Macroeconomy
20
10
0
-10
-5
0
5
Real GDP growth
-5
0
5
0
5
Trade/GDP
-5
-5
5
Domestic Credit Growth
100
50
0
-50
0
5
20
0
-20
-40
0
5
Export Growth
0
-5
Inflation
40
20
0
-20
-5
30
20
10
0
-10
Investment/GDP
200
100
0
-100
100
50
0
-50
40
30
20
10
-5
0
5
Current Account/GDP
3
2
1
0
-1
-5
0
M2 Growth
5
-5
0
5
Reserves/M2
Means and +/-2se CI; Chinese post-'99 means shown
Annual Movements around (27) Fix Exits with Appreciation
49
International Reserves
.8
1.5
.6
1
.4
.5
.2
0
0
-5
0
5
-5
Reserves/GDP
0
5
ST Debt/Reserves
15
3
10
2
5
1
0
0
-5
0
Reserves/Import Months
5
-5
0
5
Reserves/M2
Means and 90%CI; Chinese post-'99 means shown
Annual Movements around (27) Fix Exits with Appreciation
50
Growth Rates of GDP Components
30
20
20
10
10
0
0
-10
-5
0
5
Household Consumption Growth
-5
40
0
20
-50
0
-100
-20
0
Investment Growth
5
Government Consumption Growth
50
-5
0
5
-5
0
5
Import Growth
Means and 90%CI; Chinese post-'99 means shown
Annual Movements around (27) Fix Exits with Appreciation
51
Few Changes in General
• Historically, “exit-ups” unexciting
• Early days, but little reason to believe that
China’s switch will be historically atypical
52