Title: Real Convergence in Central and Eastern Europe Recent developments and some lessons of experiences
1Real Convergence in Central and Eastern Europe
Recent developments and some lessons of
experiences
Special OeNB East Jour Fixe Central and Eastern
Europe Is Economic Convergence on Track? 22
June, 2007
- Gábor Oblath
- Magyar Nemzeti Bank (NBH)
- Monetary Council
2Robert Solow (2001)
- the cross-country regression is not the only
way to make comparisons. You could intend just to
interpret parallel time series for a small group
of countries, with the goal of understanding the
source of differences among them. - (Robert Solow Applying Growth Theory across
Countries. The World Bank Economic Review, 2001,
No. 2)
3Scope and limitations
- Tried to follow Solows guidance limited success
regarding CEE-10 - (For CEE-10 cross-country regressions almost
best explanation initial income (sub-)
regional dummies) - Instead
- place recent rapid real convergence of CEE-5 into
a longer (historical) perspective - call attention to implications of alternative
statistical sources and economic indicators for
the interpretation/measurement of convergence - some factors related to convergence and some
paradoxes (SL-SK) - convergence among, and divergence within,
countries - the role of macroeconomic, in particular, fiscal,
stability Hungarys experience - Regional scope
- mainly CEE-5 (V5 CZ, HU, PL, SK, SL), but also
- CEE-8 (V5 B3 (EE, LAT, LIT)
- CEE-10 CEE8 SEE2 (BU, RO)
- Limitations
- not a formal/systematic treatment of the topic
- no detailed discussion of country specific
problems touch upon HU and PL - no treatment of effects of EU-transfers, FDI etc.
- not necessarily the most important/topical issues
4Points of departure
- For a Hungarian economist inconvenient to talk
on real convergence of CEE-10 in 2007 (HU
convergence expected to pause in 2007-8) - For a citizen of the CEE-region a pleasant
topic recent developments - CEE-8 linked to a lucky environment,
characterised by long-term real convergence
(forming a convergence club) - CEE-10 members of the EU which specifically
aims at convergence (cohesion) - Recently high-quality analyses of real
convergence of CEE-8 countries by international
institutions/ (individual scholars) - Analysis by NBH (Convergence Report, 2006)
- I agree with most of their common findings, in
particular - Labour participation/activity/employment rate
generally low in CEE ? to be increased - Investment in human, beside physical, capital
essential - Former sources of TFP-increase (reallocation,
structural change) expiring - Institutional environment essential
- Domestic savings vs. investment, FDI, CA ?
- Importance of macro-stability
- Do not wish to repeat these findings focus on
other issues and develop on some examples
/ E.g. ECB Determinants of Growth in the
CEE-EU Member States, 2007 IMF Growth in the
CEE Countries of the EU, 2006 EU Catching-Up,
Growth and Convergence of the New Member States,
2004
5Outline
- Real convergence concept, indicators, statistics
recent developments CEE-10 CEE-8 - A longer perspective than 1993-2006 is it
relevant? Pros and cons - Implications of alternative measures of real
convergence - On some factors related to per capita output
convergence - Overall convergence and within-country
divergences - The role of macroeconomic stability HU
experience, and some stability/vulnerability
issues in other CEE-countries
6Real convergence concept, indicators, statistics
- The concept of convergence increase in the
relative level of development ( decrease in
relative underdevelopment) - Getting closer to the level of more developed
economies - catching up (relative level)
- Decrease in the distance
- closing the gap
- Implication of the two interpretations?
- Indicators
- simple catch up ?
- sigma and
- beta convergence ?
- Statistics several sources often confusing ?
7Statistics different approaches and sources
(PPP essential tool of cross-country
comparisons)
- Alternative computations of PPP-s - demand side
only recently supply side - demand-side EKS and GK
- GK items additive, but bias (Gerschenkron-effect)
- EKS items non-additive, but no base-country bias
- Sources
- PEN WT (GK)
- Groningen (both GK and EKS)
- Eurostat PPS/ (EKS) We mainly rely on Eurostat
and Groningen - OECD (EKS)
- Longer series Maddison (very long) and Groningen
(since 1950) - Constant vs. current PPPs
/PPS special PPP for the EU (the purchasing
power of 1 PPS 1 euro for the EU-25)
8Statistics - two measures of differences in
growth rates vs. EU-15 (1995-06) volume indices
and PPS (in -points)
(right)
Deviation from EU-15 annual growth rate
Deviation of the two measures
Source AMECO
For most of CEE-10 countries volume indices imply
faster catch-up than PPS exception SK and
RO Current-price PPP figures meant for
cross-country comparisons in a particular year
9Convergence and divergence in per capita GDP of
CEE what is the proper point of departure?
- Three questions
- Does history begin in 1993? (or 1995, the first
year of comparable/consistent figures by
Eurostat) - Or in 1989 involving the deep recession
- Or much earlier
- state socialism (1950 -)
- or as early as data are available (Maddison)
10Per capita GDP-convergence(Constant, 2006-PPP-s,
EU-15100)
Source own calculations based on Groningen
Growth and Development Centre and the Conference
Board, Total Economy Database, January 2007
From which year to consider actual convergence?
1991, 94 or 2000? Depending on the base year
selected diverse performance of individual
countries. Some still below 89
11Per capita annual GDP growth rate differences vs.
EU-15 (Period1989-2006 moving base
years)(Which is the proper base-year?)
- - If a country has a negative growth
- rate difference on the base year 1989, it
- is still below its relative level in 1989
- In principle the longer the period of
observation, - the more relevant regarding underlying trends
- Two problems
- - interpreting transition recession
- - structural changes quality changes
12The effect of relative population and GDP growth
on GDP/cap growth vs. EU-15(1990-2006)
Population of EU-15 increased by 5.5 in CEE-10
decreased by 2.5 (B-3 -9.5)
Contribution of population and GDP change to
GDP/cap relative to the EU-15/
Source own calculations based on the Groningen
database
CEE-8 no relative GDP volume change catch-up
due to relative population-decline
/differences of log changes
131995-06 (2001-06) CEE GDP/capita growth relative
to EU-15 contribution of relative GDP growth
The picture is different
1995-2006
2001-2006
Contribution of relative decline in population
14Similarities and differences in country-patterns
Log differences in GDP/cap from EU-15 (3-year
moving averages)
Common initial U-shape, but later significant
differences
B-3 and SEE-2
V-5
SL
CZ
HU
EE
SK
PO
LI
LAT
BU
RO
Bump in the catch-up CZ, BU, RO The only
country with relatively mild initial fall and
continuous catch-up Slovenia
Source calculations based on the Groningen
database
15Regional averages(GDP/cap growth-rate
differences vs. EU-15)
16Deviations from regional averages
17Change in relative growth rates from the previous
period (1995-06)
HU different but why?
18(Relatively) longer-run relationship between per
capita GDP and growth rate differences between
EU-8 and EU-151989-2006
19Between 1995 and 2006
20Between 1999 and 2006
21Let us compare
Depending on the initial year
significant differences between initial income
and the speed of convergence across CEE-8
22Distance from EU-15 in 1989 and 2001 and relative
annual growth differentials
23Digression implications of historical statistics
- 1960/1991 - Eurostat
- 1950 - Groningen
- Long series - Maddison
- UN-ECE 1950-90
- Any implications for future catch-up?
24In most analyses of CEE relative growth rates
(real convergence), history begins in 1991-93Per
capita GDP levels at current PPS (EU-15100)
Source Eurostat
25History real divergence Per capita GDP of 3 (4)
CEE-countries relative to Austria, 1950-2006in
GK-PPP1990 (left) and in EKS-PPP2006 (right)
Which story do you believe?
EU-15
EU-15
HU
CZ_SK
CZ
HU
CZ
SK
SK
PL
PL
Source calculations based on the Groningen
database
Regarding initial relative levels for the
1950-s the left, For the 2000-s the right-hand
side diagram seems to be right. But what
happened in between? How to reconcile the
differences?
26Very long term convergence/divergence vs. the
developed West European (WE-12) countries
WE-12100
Source calculations based on Maddison
Historical Statistics for the World Economy
27UN-ECE (2000) on longer term divergence of 3 CEE
countriesEU average 100
HU
CZ-SK
PL
28Are these historical statistics relevant?
- Fundamental problems with the quality of data
- Data based on PPPs of 2006 and 1990 show totally
different picture - Not clear to what extent quality changes
accounted for in price indices - a) during centrally planned period
- b) during and after transformation-recession
- Fundamental problems of interpretation
- If sound, do these figures imply that CEE is
converging to its historical distance from WE? - Or two structural breaks
- implementing the framework of a market-economy
- joining the EU
- Imply a break with the past?
- Important message divergence of CEE from WE
started much earlier than 1989
29The meaning of real development(output vs.
income)
- Output GDP at PPP
- Per person
- Per employed person
- Per hour worked
- Income
- GDP?
- RGDI (real GDP corrected for the change in the
terms of trade implicit income transfers from/to
RoW) - GNI
- GNDI
- RGNDI
- ? RGNDI net capital transfers
Labour productivity
30PPS-GDP per capita, per person employed and per
hour worked (EU-25100)
Source Eurostat
Diverse patterns huge differences in
labour-intensity and labour utilization HU, SK
and PL similar to GR B3 and CZ similar to Spain
31Labour intensity and labour utilisation,
2005(EU25100)
High intensity, but very low labour utilisation
PO, HU (GR) Low intensity/high utilisation SI
(PT)
32PPP-GDP per capita, per worker and per hour
worked (EU-15100)
Per person employed ?
Per capita
Per hour worked
Not only levels, evolution of the three
indicators also diverse across countries
Eg CZ - SK
Source Groningen database
33Per capita RGDI change
- GDP represents output
- RGDI real income of a country (output corrected
for the impact of changes in the terms of trade) - Is it real? (are foreign trade price indices
accurate?) - Perhaps measurement problems of prices, but if
so opposite measurement problems in net exports
(volumes) RGDI useful measure - All in all if T/T show a trend, RGDI is relevant
for income growth
34Cumulative differences in RGDI and GDP growth
rates since 1995
35Cumulative difference between RGDI and GDP
growth and annual growth rate of GDP1995-2006
Percentage points
percent
Source Eurostat, AMECO
36The significance of changes in the terms of
tradenumber of years (at average GDP-growth)
corresponding to the cumulative difference
between RGDI and GDP 1995-2006
37Per capita GDP and RGDI in 2006 (EU15100)
Per capita GDP at 1995 PPS and RGDI at 1995 PPS
Per capita GDP at 2006 PPS
- Two points
- differences between constant (1995) and current
(06) PPS levels - differences between RGDI and GDP levels at prices
of 1995 -
38GDP/cap (PPS) in 1995 (x) and in 2006 RGDI in
2006 at 1995 PPS (y)(EU-15100)
2006 GDP/cap RGDI/cap
GDP/cap (PPS) 1995
39Comparison of GDP and RGDI convergence of CEE-10
1995-2006
/
/
/ log(Y06/Y95)/t where Y Yi/Yeu15
/ ?log(RGDI_06)/(GDP_06)/(Conv_GDP)
40Some factors related to convergence
- Technical
- initial level
- the depth of transformation recession
- Investments
- Export-performance
- Macroeconomic stability and policy ?
41Technical contributions to convergence 95-06
(CEE-10 and CEE-8) initial relative GDP/cap
and the fall between 1990-94
8 countries
10 countries
Log diffrence from EU-15 in 1995
Fall between 90-94
42Factors of convergence investment rate (I/GDP)
and GDP growth rate lagged (left) and
contemporaneous (right)
I/GDP
I/GDP
Apparently lagged investment has no effect on
growth (negative) growth seems to affect
investments
I/GDP
43Relative export performance (X)and relative
GDP-growth (Y) 2000-2006important for CEE
Gross domestic product at 2000 market prices
Performance relative to the rest of 35 industrial
countries (double export weights)
Exports of goods and services at 2000 prices
Performance relative to the rest of 35
industrial countries (double export weights)
44Regional issues- examplesHungary regions
Annual change in per capita GDP 1995- 2004
HU
EU-15
Log per capita GDP (PPS) in 1995
45Hungarys counties
Annual change in per capita GDP (PPS) 1995- 2004
Log per capita GDP (PPS) in 1995
46The change in counties relative position vs. the
EU-15 between 1995 and 2004
Bud
Per capita GDP (PPS) relative to EU-15 in 2004
HU
Per capita GDP (PPS) relative to EU-15 in 1995
47The Czech Republic
Per capita GDP (PPS) relative to EU-15 in 2005
Annual change in per capita GDP (PPS) 2000-2005
CZ
EU-15
Per capita GDP (PPS) relative to EU-15 in 2000
Log per capita GDP (PPS) in 2000
48Macroeconomic stability, economic policy and
institutions
- Real, price and wage convergence (the relative
level of the real exchange rate vs. real economic
development) - External balance and its domestic counterparts
- Monetary policy
- Fiscal developments
- (Institutional indicators)
49GDP/cap and price levels in 2005,European
countries (EU-15100)
CZ
GDP/cap (PPS)
50Convergence in per capita output and in the price
level of GDP (1995-2005) different
patterns(EU-15100)
Price level of GDP
Per capita GDP at PPS
51Convergence in per capita output and in the price
level of GDP (1995 2005) the Baltic and the
V-5 countries(EU-15100)
EE
SI
HU
PL
CZ
LV
SK
LT
52The above picture changes, if the price level
compared with LP (per hour)2005 EU-15100
Price level of GDP
GDP/hour worked
53Relative productivity (per hour) and relative
real product wage levels (EU-15100)
SL
Comp/emp at GDP-PPP
HU
CZ
PL
SK
Relative productivity
SK vs SL?
54Relative real producer wage increase over
relative productivity growth ?explanation
Source AMECO
55Relative productivity growth vs. relative real
product wage increase 1999-2006
Gross domestic product at 2000 market prices
per person employed - Performance relative to
the rest of 35 industrial countries double
export weights
Real compensation per employee, deflator GDP
total economy - Performance relative to the rest
of 35 industrial countries double export weights
Output growth over wage growth relative to
competitors ? change in relative profitability
56Domestic counterpart of the CA deficit
Luengnaruemitchai, Pipat-- Schadler, Susan Do
Economists' and Financial Markets' Perspectives
on the New Members of the EU Differ? IMF-WP07/65
(March 2007)
57Strong negative relationship between CA balance
and real convergence (2000-06)(larger deficit
accompanied by faster convergence 0,6 pp
increase in deficit ?1 pp increase in relative
per cap growth)
Relative growth rate
CA/GDP
But deficits above 10 may be unsustainable
58Monetary policy real short-term interest rates
in three countries
Source AMECO
59Poland real interest rate and real convergence
60Fiscal balances overall improvement except HU
HU
61Positive relationship between fiscal balance and
relative growth 2001-2006 (smaller deficit ?
faster real convergence 0,55 pp decrease in the
deficit ? 1 pp increase in per cap. relative GDP
growth)
Relative growth rate
HU
Fiscal balance/GDP
62HU real convergence (per cap. growth differences
vs. EU-15) and the fiscal balance 5 phases(the
cumulative effect of fiscal expansion on real
convergence clearly negative)
Relative growth
Fiscal balance
Temporary correction
Further stimulus/ correction negative effect
Further stimulus small effect
Temporary stimulus to growth
Correction strongly negative effect
63The fiscal impulse and difference in relative
growth rate vs. EU-15 HU 2000-2007
64Possible channels between larger deficit and
lower growth
- Uncertainty regarding the time and mode of
correction - Uncertainty regarding the exchange rate (currency
crisis?) - Negative impact on private investment
- Temporary extra boost to domestic demand during
expansion large negative impact of fiscal
correction - High risk premium, high interest rate
- high taxes, tax wedge
- Non-Keynesian effects? Political
(un-)feasibility of focusing only on the
expenditure side
65Gross fixed capital formation volume indices in
the V-5
66Tax burden of GDP
SK vs. SL the country with the
highest/increasing tax burden and the one with
the lowest/decreasing burden perform best In the
2000s.
67Institutional environment
68Finally EU-Commission forecasts for 2007-2008
(EU15100)
69Summing up no clear patterns, inconclusive
lessons no forecast
- Differences in volume indices vs. PPP measures
and GDP vs. RGDI no simple explanation - The growth/catch-up performance of CEE markedly
improved in the 2000s - On macro policies/macro environment
- Different macroeconomic/structural policies, or
different behaviour in the private sector
(institutions) ? unclear (geography important
explanation) - CA deficits help convergence, but fiscal
deficits harm convergence (implication twin
deficits harm) - excessive monetary tightening (PL) and
- excessive fiscal loosening (HU) temporarily
harm real convergence - large CA deficits may increase vulnerability of
the Baltic countries - Reforms, business environment (examples)
- Slovakia huge cut in taxes, significant reforms
- Slovenia high taxes, little change in
environment (puzzle for reform-fundamentalists)
but conservative fiscal policy strange, but
successful monetary policy - Relative prices, wages (examples)
- Slovakia relatively low wage and price level
- Slovenia relatively high wages
- But both countries recent performance and
outlook sound no simple lessons - Since 1995, in (almost) every country there were
setbacks in the convergence process - This can be expected in the future
- For HU this happens at a time when the catch-up
of all others accelerates
70Inserts
71Two ways of looking at convergencerelative
level vs. distanceCEE-8 realtive to EU-15
72Convergence of a single country example
HU(1991 vs 1995-2006)
73Alternative ways of measuring real convergence
from relative per cap PPP-GDP indices
?
74Alternative indicators of convergence (1)
- Simplest catch up rate (in )
- where
- Yi and Y indicate, respectively, the PPP-GDP/cap
of the catching up, and the target, country (or
region in our case EU-15 average) in the base
year - ? denotes absolute variation between the base and
the final year - Since, by assumption, (Yi Y) is negative, the
sign of the catch-up rate is - negative, if ?(Yi Y) is positive, i.e., the
gap is closing, and - positive, if the gap is widening
- (to get annual averages, the formula has to be
divided by t number of years between the base
and the final year)
?
75Annual catch-up rates to EU-15the early years
(-1995) and the past decade(catching up, if
index negative )
Extremely diverse patterns discussed later
76Sigma (s) and beta (ß) convergence for a group
of countries
- Sigma (s) fall in dispersion (e.g. standard
deviation in log levels) example - Beta (ß) the (negative) relation between initial
income and the rate of growth beta estimated
from Ln Yt Ln Yt-1 ? ß (Ln Yt-1)e
example - Absolute vs. conditional convergence
- absolute, if growth differences explained by
initial differences (prospect actual catch-up) - conditional, if other explanatory variables
needed
?
77(sigma)-convergence in Europe (ex. LUX)
1993-2007(Standard deviation of log per capita
GDP_PPS)Rapid fall of GDP/cap dispersion due to
convergence of low-income new members EU-14
divergence since 2002
EU-24
NMS-12 CEE-10CYMA
CEE-10
CEE-8
NMS-10
EU-14
All new members are converging, but at different
speed rapid convergence among NMS
78Alternative indicators of dispersion the same
story
Coefficient of variation
Standard deviation of Log GDP/cap
St dev of levels relative to EU-15
79Why growing dispersion since 2002 among EU-14?
(EU-15100)
Divergence of IRL() and PT and IT (-) not
compensated by convergence of GR and Spain
?
80Beta-convergence1995-2006 Log per capita GDP in
1995 vs. annual growth 95-06(slope negative ?
convergence)
CEE-8
Growth rate
26 countries
CEE-10
Log GDP/cap PPS 1995
Absolute convergence of CEE8 ? subsets
81Individual countries vs. average CEE-8