Flexing Your Muscles: Effects of Abandoning Fixed Exchange Rates for Greater Flexibility Barry Eichengreen and Andrew K. Rose UC Berkeley, NBER and CEPR
Motivation What happens when country abandons a fixed exchange rate regime, but from a position of strength, not weakness? Inverse of usual currency crisis situation Tremendous relevance: China Eichengreen Rose: Flexing 2
Empirical Orientation Identify 51 Flexings situations where country abandoned fix and could have expected its exchange rate to appreciate Empirical Question: What Happened? Short, unpretentious, ti policy relevant l paper Eichengreen Rose: Flexing 3
Answer Tremendous Heterogeneity Few apparent causes/determinants of flexes No big growth effect on average Patterns Emerge Growth falls more for countries with high investment rate Ditto rapidly growing trade Conclude: China may be right to fear the flex Eichengreen Rose: Flexing 4
Definition of Flexing (1) Shift out of Fixed Exchange Rate Use Reinhart Rogoff (with Ilzetzki) fine (15 way) exchange rate regime classification Exit from: 1) no separate legal tender; 2) preannounced peg or currency board arrangement; 3) preannounced horizontal band narrower than or equal to +/ 2%; and 4) de facto peg 1946m1 27m9, 218 countries De facto not de jure Eichengreen Rose: Flexing 5
Definition of Flexing (2) Examine Exchange Rate Changes over 3 months after exit Require appreciation or small (<5%) depreciation Motivation: reasonably expect absence of crisis Examine US$(official, black market), also SDR Ignore High Frequency Considerations Longer/Shorter Horizon doesn t matter much Changing 5% threshold doesn t matter much Review, Check each observation Eichengreen Rose: Flexing 6
51 Flexings Australia 1974 Iraq 1982 Malta 1972 Singapore 1973 Botswana198 Ireland 1979 Mauritania 1974 South Africa 1972 Canada 197 Israel 197 Mexico 1976 Spain 1974 Costa Rica1963 Italy1973 Morocco1973 SriLanka 1968 Costa Rica 1971 Jamaica 1983 Mozambique 24 Sri Lanka 199 Finland 1973 Japan 1973 Nepal 1978 Suriname 1974 France 1971 Kuwait 1975 Netherlands 1971 Sweden 1973 Germany 1969 Liberia 1998 New Zealand 1973 Switzerland 1973 Germany 1973 Libya 1971 Nicaragua 1993 Tunisia 1974 Greece 1966 Lithuania 23 Paraguay 196 Turkey 1961 Haiti 1985 Malawi 1973 Peru 1967 Turkey 1972 Hong Kong 1972 Malaysia25 Philippines 197 UK 1972 Iran 1974 Malaysia 1975 Portugal 1973 Eichengreen Rose: Flexing 7
Determinants of Flexes Causes of Exits may influence outcomes Hence must examine determinants first Little theory (Grilli); reasonable to examine over heating/inflation/asset pricebubbles Estimate Probit models, attempt to link flexes to macro/financial variables of interest Smoothed over time into 3 year averages (5 year averages similar) Modelswork poorlyinpractice practice Many variants Strong exception: size matters (intuitively) Eichengreen Rose: Flexing 8
Flex Determinants: Bivariate Panel Probits, Time FE, Country RE GDP Growth.33 M2 growth.6 Consumption/GDP.16 Reserves/M2.39 Investment/GDP.7* Reserves/GDP.27 Government spending/gdp.71 Trade/GDP.74 GDP Inflation.44 Current Account/GDP.23 Consumption growth.95 Export growth.32 Investment growth.34 Import growth.54 Government Spending growth.62 Log Population.*** Domestic Credit growth.66 Eichengreen Rose: Flexing 9
Flex Determinants: Multivariate Panel Probits, Time FE, Country RE Consumption.3 /GDP (.6) Investment.11 /GDP (.1) GDP.7 Inflation (.8) Log Population.81 (.58) Trade.2 /GDP (.3) Eichengreen Rose: Flexing 1
So, How Important is Selection Bias? Seems Not Very, at least in practice If flexes are random, can examine flexes without worrying much about selectivity Hence can proceed on to event studies Mostly WDI macro data Choose wide variety of macro/financial indicators Focus on China relevant ant characteristics (e.g., C/Y, I/Y) Guided by case studies (Have looked at others without success) Eichengreen Rose: Flexing 11
Default Event Study 2 1-1 -2 12 1 8 6 4 4 3 2 1 4 3 2 1 GDP Growth Consumption, %GDP Investment, %GDP Gov't Cons'n, %GDP 3 2 1-1 Inflation 3 5 2 1-1 -5 Consumption Growth Investment Growth 2-2 Gov't Cons'n Growth 6 6 2 4 4 1 2 2-2 -2-1 Credit Growth M2 Growth Reserves/M2 5-5 Reserves, %GDP 15 4 4 5 1 2 2 5-5 -2-4 -2-5 Trade, %GDP Current Account, %GDP Export Growth Means and +/- 2se CI; GDP series spliced. Annual Movements around (51) Exits Import Growth Eichengreen Rose: Flexing 12
Sensitivity Analysis 2-2 GDP Growth 1 4 4 8 2 2 6 4 Consumption, %GDP Investment, %GDP Gov't Cons'n, %GDP 3 2 1-1 Inflation 3 2 1-1 Consumption Growth 5 2 1-5 -1 Investment Growth Gov't Cons'n Growth 1 1 2 6 5 5 1 4 2-5 -5-1 -2 Credit Growth M2 Growth Reserves/M2 Reserves, %GDP 15 2 4 5 1 2 5-2 -5-4 -2-5 Trade, %GDP Current Account, %GDP Export Growth Means and +/- 2se CI; GDP series spliced. Import Growth Annual Movements around (32) Depreciation-Free Exits Eichengreen Rose: Flexing 13
Devaluations (Contrast) Eichengreen Rose: Flexing 14
Few Apparent Trends Flexes are not all alike Contrast to devaluations (Eichengreen et al) and currency crashes (Frankel Rose) Big message: flexes are heterogeneous Consistent with more detailed case studies But univariate event studies can mask covariation We look kfor big obvious ones, with Chinesespecific characteristics of interest Eichengreen Rose: Flexing 15
Pre Conditions of Flexes and Growth Growth Rates before/after (51) Exits and Pre-Conditions 3-year averages of WDI data 5 1 15-1 -5 Kuwait Malta Costa Nicaragua Morocco SiL Sri Lanka Rica CanadaLithuania Sweden Italy Mexico Philippines Mauritania Turkey Costa Haiti Malaysia Rica AustraliaBotswana France South Jamaica GreeceGermany Malaysia United Africa Sri Nepal Kingdom Malawi Mozambique Lanka Netherlands Finland Hong New Kong, Zealand Ireland China Israel Japan Iran, Peru Tunisia Islamic Spain Rep. Switzerland Portugal Singapore 5 1 15-1 -5 Kuwait Malta Morocco Costa Nicaragua SiL Sri Rica Lanka Turkey Mauritania Lithuania Haiti Canada Costa Mexico Philippines Nepal Italy Rica Sweden Sri Lanka Jamaica United Malaysia France Israel Malawi Malaysia Mozambique Kingdom South Germany AustraliaBotswana Africa Hong New Kong, Ireland Zealand Netherlands China Greece Tunisia Iran, Finland Islamic Peru Spain Rep. Japan Portugal Switzerland Singapore 4 6 8 1 12 Consumption (%GDP) 1 2 3 4 5 Investment (%GDP) 5 1 15-1 -5 Nicaragua Malta Costa Sri Lanka Morocco Rica Mauritania Lithuania Italy Canada Mexico Philippines United Sweden Costa Rica Australia Sri Lanka France Malaysia Malawi Mozambique Kingdom South GermanyGreece Malaysia Africa Botswana New Ireland Zealand Netherlands Finland Peru Hong Tunisia Kong, China Spain Japan Switzerland Portugal 5 1 15-1 -5 Iran, Islamic Rep. Nicaragua Malta Sri Lanka MoroccoCosta Rica Mauritania Canada Lithuania Sweden Italy Philippines Costa Mexico Mozambique United Malaysia Rica Australia Sri Lanka KingdomSouth Africa Germany GreeceMalaysia Botswana Finland Tunisia Peru New Ireland Hong Zealand Kong, China Japan Spain Iran, Islamic Rep. Switzerland Portugal 5 1 15 2-1 1 2 3 Consumption growth Investment growth Eichengreen Rose: Flexing 16
Pre Conditions, continued Growth Rates before/after (51) Exits and Pre-Conditions 3-year averages of WDI data 5 1 15-1 -5 Liberia Kuwait Malta Costa Rica Morocco Nicaragua SiL Sri Lanka Canada Lithuania Costa Haiti Sweden Philippines Mauritania Turkey Rica Mexico Sri Lanka Malaysia South Africa Australia Nepal Jamaica Malawi Israel Mozambique Botswana Malaysia Tunisia New PeruZealand Iran, Islamic Rep. Japan Switzerland Singapore 1 2 3 4 M2 growth -1 1 15 Malta Nicaragua SiL Sri Costa Morocco Lanka Rica Philippines Mexico Canada Lithuania Sweden Haiti Mauritania Italy Sri South Lanka Costa Rica Germany AustraliaBotswana Malawi Africa United Malaysia Greece Malaysia France Kingdom Finland Hong Ireland Netherlands PeruNew ZealandTunisia Kong, China Japan Spain Iran, Islamic Rep. SwitzerlandPortugal -1 1 2 3 Export Growth Mozam 5 1 15-1 -5 Kuwait Nicaragua Malta Costa Sri Morocco Lanka Rica Mauritania Turkey Lithuania Philippines Costa Mexico Haiti Canada Nepal Sweden Rica Botswana Jamaica Malawi Mozambique Sri Malaysia Australia Lanka Israel South Africa Malaysia Peru New Tunisia Zealand Iran, Islamic Japan Rep. Switzerland Singapore Liberia 5 1 15 Malta Nicaragua Sri Lanka Costa Morocco Rica Haiti Italy Canada Mauritania Lithuania Sweden Philippines Mexico Costa Rica Australia Sri Lanka United Mozambique Kingdom Malaysia France Germany Malawi Botswana South Africa Malaysia Greece Finland New Zealand Tunisia Hong Netherlands Peru Ireland Kong, China Japan Spain Iran, Is Switzerland Portugal 1 2 3 4 5 1 15 2 M2 (%GDP) Import Growth -1-5 Eichengreen Rose: Flexing 17
Some Non Trivial Covariation Particularly strong effect on change in GDP growth rates of: Investment (rates, growth) Trade (growth) Can be dt detected tdttiti statistically Eichengreen Rose: Flexing 18
Bivariate Impact on Change in GDP Growth Rates, Flexes.5.15** Consumption/GDP (.6) M2 growth ( 2.3) Investment/GDP.26*** ( 3.9) Reserves/M2. (.) Government 1.1 2.2 spending/gdp (1.3) Reserves/GDP (.4) Consumption growth.3** ( 2.5) Trade/GDP.1 (1.1) Investment growth.21*** ( 3.7) Export growth.1* ( 1.7) Government Spending growth.8 (.5) 5) Import growth.17*** ( 3.) 3) Inflation.9 (.6) Current Account/GDP Domestic Credit.55 growth (.8).2 (.2) Eichengreen Rose: Flexing 19
Pre Conditions and Inflation Inflation Rates before/after (51) Exits and Pre-Conditions 3-year averages of WDI data 1 2-1 Jamaica Israel Turkey Costa Rica Japan Philippines South Africa Australia Malawi aa Sweden Mexico Sri Morocco Lanka Lithuania Liberia CanadaNew Malaysia Zealand Singapore Sri Costa Peru Lanka Tunisia RicaMalaysia Mauritania Nepal Mozambique Haiti Iran, Switzerland Islamic Rep. Kuwait Botswana Nicaragua 5 1 15 Reserves (%M2) Malta 1 2-1 Jamaica Israel Turkey Italy Costa Rica Portugal Philippines United Finland South Spain Kingdom Africa Japan Malawi Australia aa Mexico Sweden Sri Morocco Lanka France Canada Liberia Netherlands Lithuania New Zealand Malaysia Singapore Ireland Costa Peru Sri Greece Mauritania Germany Lanka Nepal Tunisia Malta Rica Malaysia Mozambique Haiti Iran, Islamic Rep. Switzerland Kuwait Botswana Nicaragua 2 4 6 8 Reserves (%GDP) 1 2-1 Jamaica Israel Turkey Italy Portugal SpainSouth Costa Rica Australia Finland Japan Philippines United Kingdom Africa Malawi Mexico Morocco Sri Sweden Lanka a Lithuania a France Canada Liberia Netherlands New Zealand Nepal Ireland Greece Germany PeruTunisia Sri Lanka Malta Costa Mozambique Rica Malaysia Mauritania Haiti Iran, Switzerland Islamic Rep. Kuwait Botswana Nicaragua 1 2 Hong Kong, Malaysia China -1 Jamaica Israel Italy South Africa United Kingdom Australia Lithuania a Sweden Singapore Canada Netherlands Ireland New ZealandMalay Sri Lanka Mozambique Germany Malta Nepal Malaysia Haiti Botswana Nicaragua 5 1 15 2-3 -2-1 1 Trade (%GDP) Current Account (%GDP) Eichengreen Rose: Flexing 2
Any Other First Order Fears? Crisis incidence potentially important Different types of crises Different measures of each type Most deliver little Eichengreen Rose: Flexing 21
Flexes and Banking Crises Before Flexes After Flexes Total Non Crises 93 132 225 Crises 1 4 5 Total 94 136 23 Five year periods before/after flexes Test of Equality: χ 2 (1) =.9; p value =.3. Crises taken from Bordo et al.; other measures deliver similar results. Eichengreen Rose: Flexing 22
Flexes and FX Crises Before Flexes After Flexes Total Non Crises 83 123 26 Crises 11 13 24 Total 94 136 23 Five year periods before/after flexes Test of Equality: χ 2 (1) =.3; p value =.6. Crises taken from Bordo et al..; other measures deliver similar results. Eichengreen Rose: Flexing 23
Caveats Can t go too far with only 51 observations Analysis all reduced form Especially ill weak evidence of determinants of flexes Hints that credit booms precede flexes Accordingly, caution appropriate Short, policy oriented, unpretentious paper Eichengreen Rose: Flexing 24
Conclusion Little evidence of major effects after flexing Still, some evidence consistent with Chinese caution. Since 1999, for China: Investment > 4% GDP Export Growth > 2% Import Growth > 15% Chinese values are extreme for flexers! All associated with bigger post flex slowdowns Eichengreen Rose: Flexing 25