Transcript TRANSITION ECONOMIES IN SE EUROPE FRY/Serbia and Its

Measuring Performance and
Transition Progress
The Case of Western Balkans
Prof. Dr Bozidar Cerovic
Belgrade University, Faculty of Economics,
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Introduction
This presentation deals with various concepts of
economic performance measuring in transition
economies and will be exploring the impact of
initial conditions and policies on the GDP growth.
After a brief literature review of this increasingly
controversial issue, various approaches will be
checked using examples of Western Balkan
countries.
The results will support those studies that find initial
conditions predominant and policies as
endogenous
YICGG’08, Rome, Aug 18-26, Tor Vergata University
What did we know about economic
performance under transition?
• At the beginning of transition we knew that a kind
of recession will appear (named transition
recession or the U shaped curve of output).
• We were not fully aware how long it could last.
• By 2005/06 (app. 16 years after the start) a good
half of transition economies did not succeed to
achieve their pre-transition level of output.
• Moreover, both – ‘good’ and ‘bad’ runners –
substantially differ regarding their growth rates.
• What can explain such huge differences between
the countries?
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Literature survey
• Seminal papers – de Melo et al. (1996; 1997)
• Similarly, Sachs (1996).
– They taught us that policy predominantly matters (with an
increasing importance over time), although allow for
initial conditions of a country to have some influence (but
fading out over time).
– Policy variables were based on World Bank’s Composite
Liberalisation Index (CLI) or on EBRD assessments with a
general idea – the faster reforms, the better results.
• The first to question this doctrine were Aslund et al.
(1996) who found the CLI insignificant under certain
circumstances (dummy for war, FSU dummy,
shadow economy etc.)
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Literature survey
• Krueger & Ciolko (1998) were among the first to
point out that liberalisation speed could be
affected by initial condition i.e. may be treated as
an endogenous variable.
• Popov (2000) pointed out the importance of initial
conditions and found, under certain specifications,
the CLI insignificant.
• Further on new variables expressing institutions
capacity entered analyses to explain growth
differences (and appeared significant) although
Hoff & Stiglitz (2004, 2005) find that institutional
development could also be endogenous.
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Summary on literature
• The two approaches on what affects performance
predominantly:
– liberalisation/privatisation speed as a policy measure OR
– initial conditions and/or institutional level that make the
speed endogenous
• …clearly reflect the big-bang and the gradualist
position.
• In the light of these findings, we shall analyse the
Western Balkans case.
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Why Western Balkans?
Transition economies
On the eve of transition
process a big part of the
Western SE Europe was
known as the former
Yugoslavia, a country that
has dissolved by the early
nineties…
1990: Yugoslavia
YICGG’08, Rome, Aug 18-26, Tor Vergata University
What did happen?
• At the beginning of transition the country was seen as
a candidate for fast changes because of…
• …its institutional arrangement developed since the
mid-fifties (known as self-management or labour
managed economy) with many market oriented
elements.
• Moreover, we know (now) that former Yugoslav
republics exhibited significantly smaller distortions in
industrial structure and trade than other EE countries
(see: Godoy & Stiglitz, 2007)
• Finally, the transition project adopted by the last
Yugoslav government in 1989/90 seemed sound and
prospective…
YICGG’08, Rome, Aug 18-26, Tor Vergata University
What did happen?
• Despite a satisfactory outcome (inflation rate 0%,
privatisation started, stable exchange rate, growing
reserves etc.) there was a strong opposition and…
• …an angry debate over the state re-organisation was
speeding up reflecting a much deeper conceptual
differences:
– Weather the country will be transformed in line with
contemporary trends of liberalisation and transition or...
– … will it follow the old collectivist pattern, which
unsurprisingly, in absence of the former ideology, was
converted into nationalism.
• Under these circumstances the break up was inevitable
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Outcome: GDP drop…
…and some peculiarities
Slovenia
GDP in FY Republics, 1990-95 (1990=100)
Macedonia
120.0
100.0
Croatia
(war)
80.0
60.0
B&H
Serbia
(war) &
MNE
40.0
20.0
0.0
1
B&H
2
Croatia
3
Mac FYR
4
MNE
5
Serbia
6
Slovenia
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Why Western Balkans?
• We could observe different performance levels…
• …which might imply various transition policies in
Western Balkan countries.
• On the other hand, all these countries come from the
same institutional, pre-transition arrangement, which
implies many similar initial conditions, although with
different initial GDP levels.
• For these reasons the Western Balkans provides a good
laboratory for analysing the theories on economic
performance in transition economies.
• I shall start with the case of Serbia.
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Serbia of the nineties
Typical example of a wrong policy (and politics)
• Despite declaring transition
Serbia pursued a policy of
centralisartion and regulation
• Even the inherited market
independence of firms was
substantially restricted…
• Hyperinflation started at 100%
monthly rate in 1992 and in
January 1994 reached a 330
million per cent, monthly
rate!!!
• Politics and policy effect:
an excessive GDP drop
– Domestic market squeeze –
country break up (-25%)
– EU&UN bans on foreign trade
(wrong politics, -15%)
– Hyperinflation (-50%)
• No policy changes
whatsoever
– Disinflation programme 1994
unsupported by other reforms
• Entering new conflicts
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Some analysis
• During the nineties we wanted to show how much
Serbia was lagging behind the others due to a
wrong policy and politics and how much did it cost.
• I shall present some very simple OLS models that
we have developed and made use of in 1996 and in
1999 (as an analytical illustration rather than an
analytical tool).
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Model 1: performance depends on
transition progress (policy)
Regression Equation 1
GDP96/89 = 101.5 - 6.4 TRA + 0.185 TRAsq
R-sq = 44.2%
(2.26)
R-sq(adj) = 39.1%
Regression Equation 2
GDP98/89 = 145 - 10.8 TRA99 + 0.313 TRA99sq
R-sq = 44.2%
(2.31)
R-sq(adj) = 39.1%
Now, what decides on an achievable level of TRA in a country?
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Model 2: transition progress depends
on initial conditions (endogenuity)
Regression Equation 3
TRA99 = 17.8 + 0.00151 GDP89
(3.04)
R-sq = 28.7%
R-sq(adj) = 25.6%
Regression Equation 4
TRA99 = 16.7 + 0.000754 GDP89 + 6.53 MREFdummy
(1.96)
(4.86)
R-sq = 65.6%
R-sq(adj) = 62.5%
YICGG’08, Rome, Aug 18-26, Tor Vergata University
0
Uzbekistan
Ukraine
Turkmenist
Tajikistan
Slovenia
Slovak R
Russia
Rumania
Poland
Moldova
Lithuania
Latvia
Kyrgyz R
Kazak
Hungary
Georgia
Macedonia
Estonia
Czech R
Croatia
Bulgaria
Belarus
Azerbaijan
Armenia
Albania
assessments
TRANSITION PROGRESS ’99
EBRD assessments & Serbia
Transition advancement 1999
35
30
25
20
15
SRB
10
5
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Costs of the Delay
120
100
POSSIBLE
ESTIMAT.
REAL
80
60
40
20
0
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
years
IN MONEY TERMS – LOSS IN 2000 WAS AROUND
16 BILL. AND GDP AROUND US$ 11 BILL !!!
US$
YICGG’08, Rome, Aug 18-26, Tor Vergata University
SE Europe: GDP (1999)
Achieved and Possible
Balkans and Vicinity: Transition Progress
GDP p.c.: Real and Possible
120
100
80
60
40
20
Real GDP

Equ. 3 – Possible GDP

Slovenia
Rumania
Moldova
MacFYR
Hungary
FRY/Srb
Croatia
Bulgaria
Albania
0
Equ. 4 – Possible GDP

YICGG’08, Rome, Aug 18-26, Tor Vergata University
What do these results point at?
• We could remark that all Balkan countries except
SLO and ALB were beneath the estimated GDPpc
level (if transition progress was seen as endogenous).
• However, remark that they would have different
GDP-s anyway due to different inherited conditions.
• The observed lag could have been easily explained if
we had added some other variables in the model
(war, for CRO, political instability or poor
institutions for MAC and BUL, partially for ROM)
• Bad policies could be directly blamed only in the
case of Serbia (an excessive lag).
• On the other hand SLO performed better than
estimated. Why?
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Slovenian case
• SLO was ranked 6th by the EBRD Report (behind HUN,
CZR, EST, POL and SKR and slightly in front of LIT
and LAT in 1999; in 2007 it was ranked even 10th).
• This was to say that SLO was not particularly fast in
conducting the reforms; hence, according to policy
argument its performance should be poor.
• However, SLO achieved its pre-transition GDPpc as a
second in the group (after POL, in 1997/8).
• Note that this was expected following our simple model
(due to favourable initial conditions). However, it
performed even better.
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Slovenian case
• Explanation: SLO was gradually transforming its
institutions (inherited from the LM system) using their
advantages (contrary to other Balkan countries that
destroyed their institutions).
• Besides, SLO was assessed as one among those with
the best inherited institutions from pre-transition period
(measured according Campos, 1999).
• In that way, the SLO case confirms the argument on
institutions and their impact on growth (missed in our
simple model) together with other initial conditions and
is a perfect example why policies and performance are
endogenous.
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Serbian case after 2000
• After the peaceful revolution in October 2000 there was
a political will for faster changes. Comeback to a
transition path.
• First moves: liberalisation, monetary stability and sound
fiscal policy introduced, privatisation re-emerged, some
restructuring taken;
• First results (2001-02): inflation drop to 10%, growth
rate 4.5%, external debt decreased (from US$ 16 to 12
bill), reserves growth and according to the EBRD
indices it moved from 13 to 21 points.
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Slowdown
• After the enthusiastic
start a slowdown has
been remarked
• Reforms remained
uncompleted …
• …followed by a
higher inflation and
unstable growth rates
Year
EBRD
Index
2000
13.3
2001
16.7
125.6
2002
21
125.7
2003
21.6
102.9
2004
21.9
101.4
2005
23.3
106.4
2006
24.4
104.7
2007
24.7
101.2
Ind(2000=100)
157.9
185.7
Note: the slowdown coincided with the assassination of the prime
minister and with political instability (bad institutions) – three
government changes and two presidential elections (2003-2008)
YICGG’08, Rome, Aug 18-26, Tor Vergata University
0
Uzbekist
Ukraine
Turkmen
Tajikistan
Slovenia
SlovakR
Russia
Rumania
Poland
Mongolia
Moldova
Mac FRY
Hungary
Lithvania
Latvia
KyrgyzR
Kazach
Armenia
Croatia
Georgia
Estonia
MNE
CzechR
Bulgaria
B&H
Belarus
Azerbaij
Albania
TRANSITION PROGRESS ’07
EBRD assessments & Serbia
Transition progress (2007)
45
40
35
30
25
20
15
10
5
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Analysis
• In order to examine this slowdown we have developed
a set of new OLS models.
• Firstly, we wanted to define what would be an
appropriate transition advancement for the country.
• Our OLS model takes sum of the EBRD indices as a
dependent variable while independent were initial GDP
(GDP 1989), dummy for market reforms (MREF),
number of years under communism (INST1) and the
difference between black and official exchange rate
(INST5) for all transition economies (except Serbia).
We have estimated this OLS model for 1998 and 2001 i.e. in the years
that could approximately represent the state of Serbia in 2007 (due to
the previously explained Serbian delay).
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Analysis
Depenedent var:
EBRD 1998
EBRD 2001
EBRD 1998
promenljiva
Variable
konstanta
constant
GDP1989
MREF
Jednačina
Equation
11
Jednačina
Equation
2 2
prilagodeni
Adj. R (%)
statistics
DW statistika
N
Jednačina
Equation 14
Jednačina
Equation 25
Jednačina
Equation 36
34.1089
24.0105
20.9941
36.5108
25.1677
[2.4328]*
[3.6194]*
[2.1667]*
[2.0163]*
[3.4550]*
[2.5913]*
0.0008
0.0007
0.0009
0.0009
0.0007
0.0009
[0.0004]*
[0.0003]*
[0.0003]*
[0.0004]*
[0.0003]*
[0.0004]*
4.1680
3.9096
[1.9889]*
[1.7361]*
-0.2232
-0.2453
[0.0501]*
[0.0478]*
INST5
2
Jednačina
Equation
33
19.9226
INST1
R 2 (%)
EBRD 2001
-0.0287
-0.0296
[0.0094]*
[0.0097]*
32.7
58.8
43.8
31.0
65.0
44.5
25.9
54.7
38.1
24.1
61.5
39
1.79
2.24
1.62
1.91
1.52
1.68
23
23
23
23
23
23
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Analysis
• We can realise how much transition progress depends
(31-65%) on various initial conditions.
• As to the Serbian case we may conclude that the first
post-2000 government made use of the relatively
favourable initial conditions and…
• …that second and third government did not (if they
had done so the EBRD score would have been 28-29)
• Note that under normal circumstances this was
feasible around 2000 already!
• Now we wanted to estimate what would be an
appropriate GDPpc level
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Analysis
• We took GDPpc as a variable dependent on:
–
–
–
–
–
–
Initial GDPpc (GDP 1989)
Market reforms (MREF)
EBRD score (EBRD and EBRD-sq)
Inflation (INF; average, since 1994)
Budget deficit (BD; dummy: lower/higher than 5%)
First principal component of initial institutional conditions
(PC1INST; based on Campos, 1999)
• We have estimated this OLS model again for 1998 and
2001 for all transition economies since this represented
an approximate state of Serbia in 2007
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Analysis
Depenedent variable:
Variable
promenljiva
konstanta
Constant
GDP1989
MREF
EBRD 1998
GDPpc 1998
Eq 1 1
Jednačina
-4526.6730
[-3.497]*
0.5115
[4.568]*
2053.5260
[3.666]*
140.9810
[2.472]*
EBRD 1998 * EBRD 1998
Eq 2 2
Jednačina
-2823.0400
[-3.760]*
0.4900
[4.288]*
1977.8620
[3.499]*
3.0324
[2.564]*
INF prosek
BDP pc 1998
Eq 3 3 Jednačina
Eq 4 4
Jednačina
-4004.1190
-2482.4340
[-3.744]*
[-3.424]*
0.5157
0.4965
[4.153]*
[3.856]*
1942.9880
1880.3060
[3.337]*
[3.049]*
126.3236
[2.745]*
2.7141
[2.613]*
-0.4692
-0.4571
[-3.294]*
[-3.362]*
BD 1998
PC1INST
R 2 (%)
2
prilagodeni
Adj. R (%)
statistics
DW statistika
N
80.5
77.5
2.01
23
82.4
79.6
2.09
23
82.6
78.7
2.24
23
82.8
79.0
2.21
23
Eq 5 5
Jednačina
-2243.7930
[-2.236]*
0.4271
[3.017]*
1642.0290
[2.905]*
Eq 6 6
Jednačina
-2811.1250
[-3.209]*
0.4494
[3.116]*
1758.4530
[2.363]*
3.0469
[2.905]*
-0.3941
[-1.985]**
697.5795
[1.937]**
-61.2448
[-0.931]
2.9701
[2.897]*
-0.3531
[-1.787]**
645.6877
[1.899]**
85.1
79.5
2.22
23
84.7
80.2
2.16
23
YICGG’08, Rome, Aug 18-26, Tor Vergata University
Analysis
Depenedent variable:
Variable
promenljiva
konstanta
Constant
GDP1989
MREF
EBRD 2001
Eq 1 1
Jednačina
-5351.2620
[-3.918]*
0.5466
[4.624]*
2257.0560
[3.906]*
161.3945
[2.823]*
EBRD 2001 * EBRD 2001
GDPpc 2001
BDP pc 2001
Eq 2 2
Jednačina
-3306.9610
[-4.170]*
0.5215
[4.329]*
2194.3010
[3.791]*
Eq 3 3
Jednačina
-4689.9660
[-4.195]*
0.5549
[3.447]*
2131.8850
[2.803]*
142.5156
[3.070]*
Eq 4 4
Jednačina
-2891.4100
[-3.430]*
0.5332
[3.142]*
2083.0130
[2.588]*
Eq 5 5
Jednačina
-2398.3160
[-1.963]*
0.5555
[3.872]*
1942.1540
[1.939]*
Eq 6 6
Jednačina
-2688.0950
[-2.841]*
0.5630
[4.047]*
1997.8980
[2.107]*
-0.8474
[-3.436]*
2.9063
[2.791]*
-0.8223
[-3.313]*
2.7496
[2.997]*
-0.7800
[-3.328]*
-364.3267
[-0.635]
-28.8121
[-0.557]
2.7315
[3.021]*
-0.7470
[-3.246]*
-343.7720
[-0.635]
84.2
80.7
2.35
23
84.4
81.0
2.35
23
84.8
79.4
2.47
23
84.4
80.9
2.47
23
3.2947
[2.917]*
INF prosek
BD 2001
PC1INST
R 2 (%)
Adj. R 2 (%)
prilagodeni
DW statistika
statistics
N
82.1
79.2
2.14
23
82.4
79.6
2.09
23
YICGG’08, Rome, Aug 18-26, Tor Vergata University
The results of general importance
• Initial GDP is always highly and in 11/12 equations the
most significant (initial conditions, after 8 and 11 years)
• MREF highly sig. (somewhat less in Eq-s 5&6 for 2001)
• EBRD (policy) is the least significant (although
significant) except for equations 5&6
• Macroeconomic stability (INF) proved to be very
important (somewhat less for equations 5&6 in 1998),
although BD was insignificant for 2001
• Conclusion: in the whole fist decade (and beyond) initial
conditions strongly determine GDP levels, which cannot
be offset by policies (that is logical provided policies are
endogenous). This finding severely questions theories
advocating speedy reforms and the power of policies.
YICGG’08, Rome, Aug 18-26, Tor Vergata University
The results important for Serbia
• The results proved that after 2002 governments’
policies were below desirable (and achievable)
level.
• In terms of GDPpc this resulted in a loss of
around US$ 1000 p.c. (around 4000 instead of
around 5000) by the end of 2007…
• …which, in turn, confirms the importance of the
past elections in Serbia that endorsed a pro-reform
government, which will have to make use of the
disposable factors for faster advancement.
YICGG’08, Rome, Aug 18-26, Tor Vergata University
CONCLUSIONS
• Our results confirm that early studies on transition and
its impact on growth failed to properly assess initial
conditions and institutions (as put by Godoy and
Stiglitz, 2007).
• The level of initial development, strongly influence
market reforms and growth for a relatively long period.
It cannot be offset by rapid reforms since the policies
themselves appear endogenous.
• This may add to understanding of the Balkan lags –
except for Serbia (bad policy) – as well as in some of
the FSU and Asian countries except for China (a new
gradual approach)
• The findings might help in proper analysing of further
developments in transition economies and undrstanding
of their performance in a longer run.
YICGG’08, Rome, Aug 18-26, Tor Vergata University
References:
1.
2.
3.
4.
5.
6.
7.
8.
Aslund, A. Boone, P. and Johnson, S. (1996) “How to Stabilize: Lessons from
Postcommunist Countries” Brookings Papers on Economic Activity, 1, pp. 217-291
Campos, N. F. (1999) “Context is Everything: Measuring International Change in
Transition Economies” World Bank Working Paper, no. 2269
de Melo, M. Denizer, C. and Gelb, A. (1996) “From Plan to Market: Patterns of
Transition” Policy Research Working Paper no. 1564, The World Bank, Policy
Research Department
de Melo, M. Denizer, C. Gelb, A. Tenev, S. (1997) “Circumstance and Choice: The
Role of Initial Conditions and Polices in Transition Economies.” Policy Research
Working Paper no. 1866, The World Bank, Policy Research Department
Godoy, S and Stiglitz, J. E. (2007) “Growth, Initial Conditions, Law and Speed of
Privatization in Transition Countries: 11 Years Later” u: Estrin, S. Kolodko, G. W.
Uvalic, M. eds. “Transition and Beyond”, Palgrave Macmillan, Basingstoke, New
York, pp. 89-117
Hoff, K. and Stiglitz, J. E. (2004) “After the Big Bang? Obstacles to the Emergence
of the Rule of Law in Post-Communist Societies”, American Economic Review, Vol.
94, 3, pp. 753-63
Hoff, K. and Stiglitz, J. E. (2005) “The Creation of the Rule of Law and the
Legitimacy of Property Rights: The Political and Economic Consequences of a
Corrupt Privatization” NBER Working Paper, no. 11772
Krueger, G. and Ciolko, M. (1998) “A Note on Initial Conditions and Liberalization
during Transition”, Journal of Comparative Economics, Vol. 26, 4, December, pp.
718-734.
THANK YOU FOR YOUR ATTENTION
YICGG’08, Rome, Aug 18-26, Tor Vergata University