This paper is the logical continuation of my earlier article: Shock Therapy versus Gradualism: The End of the Debate

(Explaining the Magnitude of the Transformational Recession) Comparative Economic Studies, Vol. 42, No. 1, Spring 2000, pp. 1-57. The arguments of this earlier article, however, are reconsidered in

light of new research and evidence.

SHOCK THERAPY VERSUS GRADUALISM RECONSIDERED: LESSONS FROM

TRANSITION ECONOMIES AFTER 15 YEARS OF REFORMS

Vladimir Popov

GDP change in selected transition economies, 1989 = 100%

35

45

55

65

75

85

95

105

115

125

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

Poland

Slovenia

SlovakRepublic

Hungary

CzechRepublic

Romania

Bulgaria

Russia

Ukraine

CentralEurope

South Eastern Europe

FSU (Russia, Ukraine)

INITIAL LIBERALIZATION AND OUTPUT CHANGE DURING RECESSION

LIBERALISATION AND OUTPUT CHANGE IN CHINA AND RUSSIA

Fig. 3. Indices of economic freedom and GDP growth in Russia and China

70

100

130

160

190

220

250

1993 1994 1995 1996 1997 1998 1999 2000 2001

GD

P as

a %

of 1

994

1

1,5

2

2,5

3

3,5

4

GDP, 1993=100%, left scale

CHINA

RUSSIA

Index of economic freedom , right scale

GDP change in Asian economies, 1989 = 100% Fig. 1b. GDP change in Asian transition economies, 1989 = 100%

40

60

80

100

120

140

160

180

200

1989 1991 1993 1995 1997 1999 2001 2003

China

Vietnam

Central Europe

Uzbekistan

Mongolia

Turkmenistan

Kazakhstan

Russia

Kyrgyzstan

Tajikistan

Source: EBRD, World Bank.

GDP change in FSU economies, 1989 = 100% Fig. 2. GDP change in FSU economies, 1989 = 100%

25

35

45

55

65

75

85

95

105

115

1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Central Europe

Uzbekistan

Belarus

Kazakhstan

Estonia

Turkmenistan

Azerbaijan

Latvia

Lithuania

Tajikistan

Russia

Kyrgyzstan

Armenia

Ukraine

Georgia

Moldova

Russia was leading in economic liberalization, while Belarus was lagging

Private sector share in GDP, %

0

10

20

30

40

50

60

70

80

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Russian FederationUkraineBelarus

But Belarus and Uzbekistan are doing better (even though they are net importers of fuel), not to mention net exporters like

Azerbaijan, Kazakhstan, Turkmenistan GDP in 2006 as a % of 1989

TurkmenistanPoland

Estonia

Slovakia

BelarusHungary

Czech RepublicArmenia

KazakhstanAzerbaijan

RomaniaLatvia

LithuaniaCroatia

BulgariaRussia

KyrgyzstanTajikistan

UkraineGeorgia

Moldova

Uzbekistan

SloveniaAlbania

40 60 80 100 120 140 160 180

Russian growth is lagging behind that of oil exporters and some oil importers

Average annual GDP growth rates in CIS countries in 2000-07, EBRD estimates

02468

10121416

Kyr

gyz

stan

Uzb

ekis

tan

Mo

ldo

va

Ru

ssia

Geo

rgia

Ukr

ain

e

Bel

aru

s

Taj

ikis

tan

Kaz

akh

stan

Arm

enia

Tu

rkm

enis

tan

Aze

rbai

jan

INITIAL LIBERALIZATION AND OUTPUT CHANGE DURING RECOVERY

Fig. 8. Liberalization index by 1995 and performance in 1994-98

R2 = 0,3345

65

75

85

95

105

115

125

135

145

155

0,5 1 1,5 2 2,5 3 3,5 4 4,5

Cumulative liberalization index, 1995

1998

GD

P as

a %

of 1

994

GD

P

CZHU

POLSLOVAK

SLOVEN

VIETN

CROAT

MACED

CHINALITH

CIS countries

BULG

ESTALB

LATV

MONG

ROM

LIBERALIZATION STOCK AND OUTPUT CHANGE DURING RECOVERY

Fig. 9. Liberalisation and output change in transition economies in 1995-2003

R2 = 0,0618

R2 = 0,0229

80

100

120

140

160

180

200

1 1,5 2 2,5 3 3,5 4 4,5

Cumulative liberalisation index

2003

GD

P as

a %

of 1

995

LIBER_94LIBER2004

LIBERALIZATION STOCK AND OUTPUT CHANGE DURING RECOVERY

Fig. 10. Liberalisation increase and output change in transition economies in 1995-2003

R2 = 0,0645

80

100

120

140

160

180

200

-1 -0,5 0 0,5 1 1,5 2 2,5

Increase in cumulative liberalisation index in 1995-2003

2003

GD

P as

a %

of 1

995

Does liberalization matter? •  Vietnam and China are similar in initial conditions and in

transition results (immediate growth of output without transformational recession) despite different reform strategies:

•  Chinese reforms are the classical example of gradualism •  Vietnamese reformers introduced shock therapy treatment

(instant deregulation of most prices and introduction of convertibility of dong) in 1989

•  Differing performance of the former Soviet Union (FSU) states: •  Baltic states are the champions of liberalization and

stabilization in the region. In the Baltics, however, output fell in the early 1990s by 36-60% and even in 2005, 10 years after the bottom of the recession was reached, was still below the pre-recession maximum.

•  Uzbekistan is commonly perceived to be one of the worst procrastinators. However in Uzbekistan the reduction of output in 1990-95 totaled only 18% and the economy started to grow again in 1996

•  By 2005 only two former Soviet republics - Uzbekistan and Turkmenistan - surpassed the pre-recession level of 1989

Reasons for recession: (1) Supply shock - change in relative prices

Fig. 4. Change in relative prices and output in 1990-98 in Russian industry

R2 = 0,4062

0

30

60

90

40% 60% 80% 100% 120% 140% 160%

Ratio of 1998 proices to 1990 prices as a % of industrial average

Janu

ary

1998

out

put a

s a

% o

f Ja

nuar

y 19

90 o

utpu

t

Electric energy

Fuel

Non-ferrous metals

Light

FoodWood

Constr. materialsMachinery

Chemicals

Petrochemicals

Steel

Industrial structure and external trade distortions

Fig. 7. Aggregate distortions in industrial structure and external trade before transition and GDP change during transition

20

50

80

110

140

170

200

0 10 20 30 40 50 60 70 80 90

Distortions in industrial structure and external trade, % of GDP

1996

GDP

as

a %

of 1

998

GDP

CHINA

VIETN

SLOVEN

HUNG

MACED

CZECH

CROATEST

ARM

Reasons for recession: (2) Collapse of state institutions - government revenues and

expenditure fell in most transition economies Fig. 13. Consolidated government revenues as a % of GDP

10

15

20

25

30

35

40

45

50

55

1989 1990 1991 1992 1993 1994 1995 1996

%

Central Europeancountries

South East Europecountries

Baltic states

Central Asiancountries

Caucasian states

-China

RUSSIA

Three major patterns of change in government expenditures

Fig. 10. Government expenditure , % of GDP

0

10

20

30

40

50

60

1985 1989 1995 1978 1985 1994 1989 1996

Debt serviceDefenceSubsidiesInvestment"Ordinary government"

POLANDCHINA

USSR/ RUSSIA

The decline in government revenues is correlated with performance

Fig. 12. Change in GDP and in the share of government revenues in GDP

30

70

110

150

190

-10 -5 0 5 10 15 20 25 30 35Decline in the share of government revenues in GDP from 1989-91 to 1993-96, p.p.

VIETN

ARM

CHINA

Impact of initial conditions, institutions, liberalization:1989-96

Table 1. Regression of change in GDP in 1989-96 on initial conditions, policy factors, and rule of lawand democracy indices, robust estimatesDependent variable = log (1996 GDP as a % of 1989 GDP)For China - all indicators are for the period of 1979-86 or similarEquations, Number ofObservations / Variables

1,N=28

2,N=28

3,N=28

4,N=28

5,N=28

6,N=28

7,N=28

Constant 5.3*** 5.4*** 5.2*** 5.4*** 5.4*** 5.5*** 5.7***Distortions, % of GDPa -.005** -.005** -.003 -.006** -.007*** -.007*** -.007***1987 PPP GDP per capita, % of theUS level

-.009** -.006* -.007** -.007** -.009*** -.008***

-.008***

War dummyb -.19c -.36*** -.37*** -.45***Decline in government revenues asa % of GDP from 1989-91 to 1993-96

-.011*** -.011***

-.011***

Liberalization index .05 -.02 .03Log (Inflation, % a year, 1990-95,geometric average)

-.16*** -.20*** -.18*** -.17*** -.13*** -.13*** -.14***

Rule of law index, average for1989-97, %

.008***

Democracy index, average for1990-98, %

-.005***

-.003**

Ratio of the rule of law todemocracy index

.07*** .07*** .06*** .05*** .05***

Adjusted R2, % 82 83 83 85 91 91 90*, **, *** - Significant at 1, 5 and 10% level respectively.aCumulative measure of distortions as a % of GDP equal to the sum of defense expenditure (minus 3%regarded as the 'normal' level), deviations in industrial structure and trade openness from the 'normal' level,the share of heavily distorted trade (among the FSU republics) and lightly distorted trade (with socialistcountries) taken with a 33% weight – see (Popov, 2000) for details.bEquals 1 for Armenia, Azerbaijan, Croatia, Georgia, Macedonia, and Tajikistan and 0 for all othercountries.c Significant at 13% level.

Evaluating the impact of non-policy and policy factors:1989-98

DDeeppeennddeenntt vvaarriiaabbllee == lloogg ((11999988 GGDDPP aass aa %% ooff 11998899 GGDDPP)) FFoorr CChhiinnaa -- aallll iinnddiiccaattoorrss aarree ffoorr tthhee ppeerriioodd ooff 11997799--8888 oorr ssiimmiillaarr

EEqquuaattiioonnss,, NNuummbbeerr ooff oobbsseerrvvaattiioonnss // VVaarriiaabblleess

11,, NN==2288

22,, NN==2288

33,, NN==2288

44,, NN==2288

55,, NN==1177

66,, NN==1177

77,, NN==1177

88,, NN==1177

CCoonnssttaanntt 55..3300 44..8888 55..6688 55..7733 55..7744 55..4433 55..8866 66..0088 DDiissttoorrttiioonnss,, %% ooff GGDDPPaa --..0011 --..0011 --..0011 --..0011 --..0011 --..0011 cc ((--..0000)) ((--..0000)) 11998877 PPPPPP GGDDPP ppeerr ccaappiittaa,, %% ooff tthhee UUSS lleevveell

--..0011 --..0022 --..0011 --..0011 --..0011 --..0022 --..0011cc ((--..0011))

WWaarr dduummmmyybb --..6677 --..5588 --..3388 --..3377 DDeecclliinnee iinn ggoovveerrnnmmeenntt rreevveennuueess aass aa %% ooff GGDDPP ffrroomm 11998899--9911 ttoo 11999933--9966

--..0022 --..0011 --..0011 --..0011

LLiibbeerraalliizzaattiioonn iinnddeexx ((..1111)) ((--..0011)) ((..1111)) ((--..0066)) LLoogg ((IInnffllaattiioonn,, %% aa yyeeaarr,, 11999900--9955,, ggeeoommeettrriicc aavveerraaggee))

--..1155 --..1155 --..1133 --..1166

SShhaaddooww eeccoonnoommyy aass aa %% ooff GGDDPP iinn 11999944 --..0022 --..0022 --..0011 --..0022 AAddjjuusstteedd RR22,, %% 6677 6699 8800 8800 7722 7733 8822 8811

aaCCuummuullaattiivvee mmeeaassuurree ooff ddiissttoorrttiioonnss aass aa %% ooff GGDDPP eeqquuaall ttoo tthhee ssuumm ooff ddeeffeennssee eexxppeennddiittuurree ((mmiinnuuss 33%% rreeggaarrddeedd aass tthhee ''nnoorrmmaall'' lleevveell)),, ddeevviiaattiioonnss iinn iinndduussttrriiaall ssttrruuccttuurree aanndd ttrraaddee ooppeennnneessss ffrroomm tthhee ''nnoorrmmaall'' lleevveell,, tthhee sshhaarree ooff hheeaavviillyy ddiissttoorrtteedd ttrraaddee ((aammoonngg tthhee FFSSUU rreeppuubblliiccss)) aanndd lliigghhttllyy ddiissttoorrtteedd ttrraaddee ((wwiitthh ssoocciiaalliisstt ccoouunnttrriieess)) ttaakkeenn wwiitthh aa 3333%% wweeiigghhtt ((sseeee AAppppeennddiixx ffoorr ddeettaaiillss)).. bbEEqquuaallss 11 ffoorr AArrmmeenniiaa,, AAzzeerrbbaaiijjaann,, CCrrooaattiiaa,, GGeeoorrggiiaa,, MMaacceeddoonniiaa,, aanndd TTaajjiikkiissttaann aanndd 00 ffoorr aallll ootthheerr ccoouunnttrriieess.. ccSSiiggnniiffiiccaanntt aatt 88%% lleevveell..

Best performance: low distortions, strong institutions Worst performance: high distortions, weak institutions

INITIAL CONDITIONS (DISTORTIONS) AND INSTITUTIONS –CLASSIFICATION OF COUNTRIES

DISTORTIONSINSTITUTIONAL CAPACITY

LOW HIGH

HIGH CHINA,VIETNAM

EASTERNEUROPE

LOW ALBANIA,MONGOLIA

FSU

Explaining the magnitude of recession:

•  Differences in performance during transition depend strongly on the initial conditions:

•  The higher the distortions (militarization, over-industrialization, "under-openness" of the economy and the share of perverted trade flows), the worse is the performance

•  The higher was GDP per capita before transition, the greater were distortions embodied in fixed capital stock, the more difficult it was to overcome these distortions to achieve growth

Explaining the magnitude of recession:

•  Macroeconomic stability matters a great deal for economic performance

•  Liberalization index does not appear to be important

•  Changes in the institutional capabilities of the state have dramatic impact on performance

By the end of the 1990s most countries were already recovering from recession

TTaabbllee 66.. RReeggrreessssiioonn ooff cchhaannggee iinn GGDDPP iinn 11999944--9988 oonn nnoonn--ppoolliiccyy aanndd ppoolliiccyy--rreellaatteedd ffaaccttoorrss ((aallll ccooeeffffiicciieennttss aarree ssiiggnniiffiiccaanntt aatt 1155%% lleevveell eexxcceepptt tthhoossee iinn bbrraacckkeettss)) DDeeppeennddeenntt vvaarriiaabbllee == lloogg ((11999988 GGDDPP aass aa %% ooff 11999944 GGDDPP)) FFoorr CChhiinnaa -- aallll iinnddiiccaattoorrss aarree ffoorr tthhee ppeerriioodd ooff 11998844--8888 oorr ssiimmiillaarr EEqquuaattiioonnss,, NNuummbbeerr ooff OObbsseerrvvaattiioonnss // VVaarriiaabblleess

11,, NN==2288

22,, NN==2288

33,, NN==2288

44,, NN==2288

55,, NN==2288

CCoonnssttaanntt 44..5511 44..2255 44..5566 44..3322 44..6600 DDiissttoorrttiioonnss,, %% ooff GGDDPPaa ..000044 ..000055 ..000033 ..000033 11998877 PPPPPP GGDDPP ppeerr ccaappiittaa,, %% ooff tthhee UUSS lleevveell WWaarr dduummmmyybb ..1155 DDeecclliinnee iinn ggoovveerrnnmmeenntt rreevveennuueess aass aa %% ooff GGDDPP ffrroomm 11998899--9911 ttoo 11999933--9966

--..000033 cc --..000044

LLiibbeerraalliizzaattiioonn iinnddeexx ..0077 ..1122 ..0099 ..1100 ..0077 LLoogg ((IInnffllaattiioonn,, %% aa yyeeaarr,, 11999900--9955,, GGeeoommeettrriicc aavveerraaggee))

--..0066 ..0044

AAddjjuusstteedd RR22,, %% 2211 2277 3377 2299 3333 aaCCuummuullaattiivvee mmeeaassuurree ooff ddiissttoorrttiioonnss aass aa %% ooff GGDDPP eeqquuaall ttoo tthhee ssuumm ooff ddeeffeennssee eexxppeennddiittuurree ((mmiinnuuss 33%% rreeggaarrddeedd aass tthhee ''nnoorrmmaall'' lleevveell)),, ddeevviiaattiioonnss iinn iinndduussttrriiaall ssttrruuccttuurree aanndd ttrraaddee ooppeennnneessss ffrroomm tthhee ''nnoorrmmaall'' lleevveell,, tthhee sshhaarree ooff hheeaavviillyy ddiissttoorrtteedd ttrraaddee ((aammoonngg tthhee FFSSUU rreeppuubblliiccss)) aanndd lliigghhttllyy ddiissttoorrtteedd ttrraaddee ((wwiitthh ssoocciiaalliisstt ccoouunnttrriieess)) ttaakkeenn wwiitthh aa 3333%% wweeiigghhtt ((sseeee AAppppeennddiixx ffoorr ddeettaaiillss)).. bbEEqquuaallss 11 ffoorr AArrmmeenniiaa,, AAzzeerrbbaaiijjaann,, CCrrooaattiiaa,, GGeeoorrggiiaa,, MMaacceeddoonniiaa,, aanndd TTaajjiikkiissttaann aanndd 00 ffoorr aallll ootthheerr ccoouunnttrriieess.. c Significant at 21% level.

Impact of initial conditions, institutions, liberalization

Table 2. Regression of change in GDP in 1995-2003 on initial conditions, institutional capacity,liberalization and rule of law and democracy indices, robust estimatesDependent variable = 2003 GDP as a % of 1995 GDPFor China - all indicators are for the period of 1979-86 or similarEquations, Number ofObservations / Variables

1,N=28

2,N=28

3,N=28

4,N=28

5,N=28

Constant 105*** 91*** 99*** 78*** 99***1996 GDP as a % of 1989 GDP .33*** .45*** .46*** .24**1987 PPP GDP per capita, % of the US levelWar dummya 22.9** 42.3*** 32.0*** 19.4*Liberalization index in 1995 -19.9***Increase in the liberalization index in 1995-2003 15.3

***16.7***

17.6*** 17.6***

Decline in government revenues as a % of GDP from1989-91 to 1993-96a

-.8***

Rule of law index, average for 1989-97, % .8** 1.0*** 1.2***Democracy index, average for 1990-98, % -.6*** -.8***Adjusted R2, % 25 38 45 52 55*, **, *** - Significant at 1, 5 and 10% level respectively.aEquals 1 for Armenia, Azerbaijan, Croatia, Georgia, Macedonia, and Tajikistan and 0 for all othercountries.

OBJECTION: Speed and extent of liberalization may be endogenous

Fig. 3. Liberalization indices and distortions in industrial structure and trade patterns

012345

0 20 40 60 80Distortions in the industrial structure and trade patterns

(% of GDP)

Cum

ulat

ive

iber

aliz

atio

n in

dice

s by

199

5

CIS

Economic liberalization and democratization go hand in hand

Fig. 4. Democracy index (1990-98, average) and economic liberalization index by 1995

0

1

2

3

4

5

0 20 40 60 80 100

Democracy index (Freedom House)

Cum

ulat

ive

iber

aliz

atio

n in

dex

by 1

995

CIS and BALKANSVIETN

CHINA

CROAT, MACED

Instrumenting liberalization stock with democracy level variable: 1989-96

Table 3. 2SLS robust estimates – regression of change in GDP in 1989-96 on initial conditions,institutional capacity, liberalization and rule of law and democracy indices (Liberalization indexinstrumented with the democracy level variable)Dependent variable = Log (1996 GDP as a % of 1989 GDP)For China - all indicators are for the period of 1979-86 or similarEquations, Number ofObservations / Variables

1,N=28

2,N=28

3,N=17

4,N=17

Constant 6.4*** 6.3*** 6.0*** 6.0***Pre-transition distortions, % of GDP -.01*** -.02*** -.0041987 PPP GDP per capita, % of the US level -.007** -.01***War dummya -.45*** -.29b

Liberalization index in 1995 -.18** -.39* -.19*** -.19***Decline in government revenues as a % of GDP from 1989-91 to1993-96

-.02*** -.02***

Log (Inflation, % a year, 1990-95, geometric average) -1.7*** -.22*** -.22*** -.19***Rule of law index, average for 1989-97, % -.01c

Increase in the share of shadow economy in GDP in 1989-94, p.p. -.02*** -.015***R2, % 86 77 88 90*, **, *** - significant at 1, 5 and 10% level respectively.aEquals 1 for Armenia, Azerbaijan, Croatia, Georgia, Macedonia, and Tajikistan and 0 for all othercountries.b Significant at 12% level.c Significant at 16% level.

Instrumenting liberalization change with liberalization stock and FSU dummy variables:1995-2003

Table 4. 2SLS robust estimates – regression of change in GDP in 1995-2003 oninitial conditions, institutional capacity, liberalization and rule of law anddemocracy indicesDependent variable = 2003 GDP as a % of 1995 GDPFor China the indicator is for the period 10 years earlier.Equations, Number ofObservations / Variables

1,N=28

2,N=28

3,N=28

4,N = 28

Instruments for liberalization change in 1995-03 variable LIBER95 FSU LIBER95and FSU

LIBER95and FSU

Constant 97.8*** 95.8*** 97.7*** 79.5***1996 GDP as a % of 1989 GDP .18*War dummya 19.5* 19.8** 19.5* 25.0**Increase in liberalization index in 1995-2003 18.2*** 19.2** 18.3*** 22.9***Decline in government revenues as a % of GDP from 1989-91 to 1993-96

-.76*** -.78** -.76*** -.65***

Rule of law index, average for 1989-97, % 1.24*** 1.28*** 1.25*** 1.13***Democracy index, average for 1990-98, % -.76*** -.76*** -.76*** -.62***R2, % 55 54 55 56*, **, *** - significant at 1, 5 and 10% level respectively.aEquals 1 for Armenia, Azerbaijan, Croatia, Georgia, Macedonia, and Tajikistan and 0 for all othercountries.

Conclusions

•  The impact of the speed of liberalization at the initial stage of transition, i.e. during the transformational recession, appears to be negative, if any.

•  The reason for the negative impact is most probably associated with limited ability of the economy to adjust to new price ratios

Conclusions

Fig. 5. Hypothetical trajectories of output (Year "0" = 100%) assuming gradual and instant liberalization

80%85%90%95%

100%105%110%115%120%125%

0 1 2 3 4 5 6 7 8 9 10

Years

Out

put,

Year

"0"

= 1

00%

Reduction ofoutput in aNC sector at10% annually

Reduction ofoutput in aNC sector at30% annually

Reduction ofoutput in aNC sector at100%annually

Assumptions: size of non-competitive sector (NC) in the initial year = 20% of total output; net investment (s) = 10% of total output;

marginal capital productivity, output increase per unit of net investment (a) = 1/3.

Conclusions •  At the recovery stage liberalization starts to affect

growth positively, whereas the impact of pre-transition distortions disappears. Institutional capacity and macroeconomic policy continue to be important prerequisites for successful performance.

•  Liberalization at the recovery stage influences performance positively because it creates market stimuli without causing rapid collapse of output of ineff icient industr ies , which cannot be compensated fully by the rise of efficient industries due to investment constraints.